M E D I C A L S U R V E I L L A N C E M O N T H L Y R E P O R T

MSMR

A publication of the Armed Forces Health Surveillance Branch

JUNE 2018

Volume 25Number 6

P A G E 2 Polypharmacy involving opioid, psychotropic, and central nervous system depressant medications, period prevalence and association with suicidal ideation, active component, U.S. Armed Forces, 2016Richard P. Eide III, MD, MPH; Shauna Stahlman, PhD, MPH

P A G E 1 0 Variations in the incidence and burden of illnesses and injuries among non-retiree service members in the earliest, middle, and last 6 months of their careers, active component, U.S. Armed Forces, 2000–2015Colby C. Uptegraft, MD; Shauna Stahlman, PhD, MPH

P A G E 1 8 Diagnoses of eating disorders, active component service members, U.S. Armed Forces, 2013–2017Valerie F. Williams, MA, MS; Shauna Stahlman, PhD, MPH; Stephen B. Taubman, PhD

P A G E 2 6 Brief report: Department of Defense midseason vaccine effectiveness estimates for the 2017–2018 influenza seasonLisa Shoubaki, MPH; Angelia A. Eick-Cost, PhD, ScM; Anthony W. Hawksworth, BS; Zheng Hu, MS; LeeAnne Lynch, MPH; Christopher A. Myers, PhD; Susan Federinko, MD, MPH

P A G E 2 9 Letter to the Editor

NIAID

CE/CME

MSMR Vol. 25 No. 6 June 2018 Page 2

This report uses routinely collected data in the Defense Medical Surveillance System (DMSS) to explore the period prevalence of polypharmacy among the active component U.S. military in 2016. The period prevalence across the Department of Defense was 10.8% and was highest for the Army (14.5%) and lowest for the Marine Corps (7.4%). Furthermore, a case control study was conducted to explore the potential association between polypharmacy and incident suicidal ideation (SI). There was an increased adjusted odds of inci-dent SI within 12 months following polypharmacy exposure, with adjusted odds ratios ranging from 1.53 (95% CI, 1.38–1.71) to 3.06 (95% CI, 2.00–4.70), depending on the number of qualifying polypharmacy criteria. Impor-tant limitations to the current analysis are discussed. Results suggest that it would be prudent to screen for SI during the polypharmacy clinical encoun-ter, particularly for persons with any of the mental health disorders consid-ered in this report. Inclusion of Department of Defense Suicide Event Report (DoDSER) data along with medically coded SI in future surveillance would increase the sensitivity of identifying incident cases of SI.

Polypharmacy Involving Opioid, Psychotropic, and Central Nervous System Depressant Medications, Period Prevalence and Association with Suicidal Ideation, Active Component, U.S. Armed Forces, 2016Richard P. Eide III, MD, MPH (MAJ, USA); Shauna Stahlman, PhD, MPH

Polypharmacy is an ill-defined term with at least 24 unique definitions in the medical literature.1,2 For exam-

ple, polypharmacy can refer to the simulta-neous use of multiple prescription drugs to treat an individual for one or more medical conditions. Polypharmacy also can be used to describe an individual’s pattern of exces-sive healthcare utilization for the purpose of obtaining prescription drugs. Polypharmacy poses the potential threat of harm from the cumulative impact of drug effects on vari-ous human tissues, interactions between one or more drugs, side effects, or a combi-nation of all these. Between 1999 and 2012, the prevalence of polypharmacy increased from 8.4% to an estimated 15% of U.S. adults aged 20 years and older.3 Concerns about

polypharmacy have historically focused on adults ≥65 years of age4-10; however, recent research shows that polypharmacy among younger adults, such as the active duty mili-tary where the overall average age in 2016 was 28.5 years,11 is a growing problem.12-14 According to a cross-sectional analysis of 311,400 Veterans Health Administration (VHA) beneficiaries during 2010–2011, polypharmacy affected 8.4% of Operation Enduring Freedom (OEF) and Operation Iraqi Freedom (OIF) veterans.15

For the purposes of this report, poly-pharmacy was defined (see Methods for complete definition) according to Defense Health Agency16 guidance used to screen Military Health System (MHS) beneficia-ries with potentially high-risk medication

use to target interventions (e.g., referral to a clinical pharmacist).17 This guidance was derived from the 2015 U.S. Army Office of the Surgeon General (OTSG) Policy Memo 15-039,18 which focused on polypharmacy associated with psychotropic drugs and cen-tral nervous system depressants (CNSDs) from seven broad categories. Psychotro-pic medications (e.g., stimulants, antide-pressants, antipsychotics) act directly on the central nervous system to affect mood, cognition, or perception; CNSDs (opioids, anxiolytics, anticonvulsants, and hypnotics [sleep aids]) down-regulate central nervous system function and some have the poten-tial to suppress function of the respiratory center in the brainstem. The use of multiple psychotropic and/or CNSD medications, especially with the addition of an opioid, is a type of polypharmacy with potential for multiple adverse events19,20 including overdose,21,22 which can result in delirium, respiratory suppression, or death.23-26 Often these adverse events are unintentional27-31; however, the authors of the previously cited VHA study found that veterans with polypharmacy (defined as five or more concurrent prescriptions) had odds of sui-cide-related behavior that were nearly four times higher (adjusted odds ratio [AOR] 3.94; 99% CI, 3.58–4.33) than veterans with-out polypharmacy after controlling for exist-ing mental illness.15

Suicide is one of the top 10 leading causes of death in the U.S.32 and the overall age-adjusted rate of suicide increased 24% between 1999 and 2014.33 Since the outset of the war in Afghanistan in 2001, the Depart-ment of Defense (DoD) has experienced a dramatic rise in suicide rates.34 The Army sustained the largest increase in suicide rate, from 8.7 per 100,000 persons in 2001 to 24.4

This article provides continuing education (CE) and continuing medical education (CME) credit. Please see information at the end of the article.CE/CME

June 2018 Vol. 25 No. 6 MSMR Page 3

per 100,000 persons in 2015,35,36 surpassing the age- and sex-adjusted U.S. civilian rate for the first time in 200835 and only recently declining to a level comparable to the civilian rate.36 Over the same period, the DoD has made significant efforts to better understand suicide within the ranks to inform policy and develop a comprehensive strategy to combat and prevent suicide. For example, the DoD developed a standardized web-based report-ing tool (DoD Suicide Event Report [DoD-SER]),37 which has been in use since 2008. In addition, the services have conducted large-scale research projects, including collabora-tion with the National Institutes of Health, as was done for the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS).38 Furthermore, in 2011, the Defense Suicide Prevention Office39 was established to lead and coordinate preven-tion activities across the DoD. As a result of these efforts, many characteristics of fatal40-42 and non-fatal43-45 suicidal behaviors among U.S. military personnel over the past 16 years have been described.

Suicidal ideation (SI) is considered an important outcome that can predict more serious suicidal behavior.43,44,46 Life-time prevalence of SI has been estimated at approximately 14% among active compo-nent Army personnel.43 Recent analysis of more than 10,000 cases of medically docu-mented SI among active component Army personnel between 2006 and 2009 found that the highest risks of SI were associated with enlisted service members with less than 2 years of service, females, and those with a recent mental health diagnosis.47 Addition-ally, an examination of 2015 DoDSER data by the Army Public Health Center found that 13% of SI cases met criteria for any polypharmacy (using the same definition from OTSG Policy Memo 15-039) at the time of reporting.48

The current level of polypharmacy across the DoD is not well established. This report examines the prevalence of polyphar-macy among active component U.S. service members in 2016 and explores the potential association between polypharmacy involv-ing opioids, and other drugs affecting the central nervous system, and SI that is inde-pendent of existing psychiatric illness and military and demographic characteristics.

M E T H O D S

All data used to identify polypharmacy and SI were derived from records routinely maintained in the Defense Medical Surveil-lance System (DMSS). These records include both ambulatory encounters and hospital-izations of active component members of the U.S. Armed Forces in military and civilian (if reimbursed through the MHS) treatment facilities. In addition, these data contain administrative records for all prescriptions written for service members at military treat-ment facilities or through civilian purchased care. For the purpose of these analyses, pre-scriptions were limited to one for each drug name per day per service member.

Prevalence of polypharmacy in 2016 was based on the referent population of all individuals serving in the active com-ponent on 30 June 2016. The surveillance period was 1 January 2016 through 31 December 2016. The surveillance popula-tion included all individuals who served in an active component of the Army, Navy, Air Force, or Marine Corps who were in ser-vice on 30 June 2016. Prescription drug data in DMSS were grouped into the seven cat-egories according to U.S. Food and Drug Administration–established pharmaco-logic class and included opioids, stimulants,

antidepressants, antipsychotics, anxiolytics (including mood stabilizers), anticonvul-sants (including muscle relaxants), and hyp-notics.49 Polypharmacy was derived from OTSG Policy Memo 15-03918 (Figure) and defined using the following criteria: Group A: four or more prescriptions all within 30 days of one another for any drug (including antibiotics, antihypertensives, statins, etc.) where at least one is an opioid; or Group B: four or more prescriptions all within 30 days of one another from among the seven cat-egories of psychotropic or CNSD medica-tions listed above; or Group C: three or more emergency room (ER) visits during the sur-veillance period, each within 7 days of an opioid prescription. Of note, for efficiency, the group C inclusion criteria deviate from OTSG Policy Memo 15-039 by limiting anal-ysis to the surveillance period (calendar year 2016) rather than any consecutive 12-month interval that has case-defining (e.g., third) ER visit associated with an opioid prescrip-tion in 2016. Additionally, persons who met polypharmacy criteria were further grouped into strata consisting of those meeting cri-teria for only one group as “low” (i.e., only group A, B, or C), those meeting criteria for two groups but not three as “moderate” (e.g., groups A and B but not C), and those meet-ing criteria for all three groups as “high.”

Prior incident mental health diagno-ses were defined according to Armed Forces

F I G U R E . Criteria for types of polypharmacy, adapted from U.S. Army Office of the Surgeon General Policy Memo 15-03918

MSMR Vol. 25 No. 6 June 2018 Page 4

Health Surveillance Branch standardized health surveillance case definitions50 and included: adjustment disorders, alcohol-related disorders, anxiety disorders, bipolar disorders, depressive disorders, personality disorders, psychoses, post-traumatic stress disorder, schizophrenia, and substance-related disorders. For the period prevalence calculation, individuals were considered to have an existing mental health diagnosis if it was diagnosed before the end of the surveil-lance period (31 December 2016).

A matched case-control design was used to explore the association between inci-dent SI that occurred during 2016 and poly-pharmacy in the preceding 12 months. The surveillance period for this component of the analysis was, therefore, 1 January 2015 through 31 December 2016. The surveillance population included all individuals who served in an active component of the Army, Navy, Air Force, or Marine Corps at any time during 2016 with evidence in DMSS of at least one existing mental health condition (defined above). An incident case of SI was defined as one inpatient encounter with a diagnosis of SI (ICD-10: R45.851) in the first or second diagnostic position, two outpatient medical encounters within 180 days of each other with a diagnosis of SI in the first or sec-ond diagnostic position, or one outpatient medical encounter in a psychiatric or men-tal healthcare specialty setting with a diag-nosis of SI in the first or second diagnostic position. Individuals who received a diagno-sis of SI before 1 January 2016 were excluded as prevalent cases. Each individual could be counted as an incident case once during the surveillance period. Controls were sampled cumulatively from those in service at the time of the SI case diagnosis, assigned at a ratio of three controls per case (3:1), matched on age (within 1 year) and sex. Cases and controls were excluded if the service mem-ber deployed within 12 months prior to the incident date of SI diagnosis. Analysis of matched groups was conducted by using conditional logistic regression to estimate the odds of SI among those with polypharmacy, compared to those without after adjusting for potential confounders, including race/ethnicity, military rank/grade, deployment history, marital status, education level, length of service, and military occupation.

R E S U L T S

Period prevalence of polypharmacy

The percentages of polypharmacy over-all and stratified by covariate subgroups are shown in Table 1. In 2016, a total of 139,249 individuals met any of the three criteria for polypharmacy, corresponding to an overall period prevalence of 10.8% across the active component military. The highest levels of polypharmacy were among Army members (14.5%) and the lowest levels were among those in the Navy (7.8%) or the Marine Corps (7.4%). The most common type of polypharmacy in 2016 was group A (those with four or more prescriptions, at least one opioid), accounting for 120,569 (86.6%) of the individuals. The highest-risk group (those who met all three criteria) accounted for only 976 (0.7%) of all individuals with polypharmacy. Overall, 54,860 (39.4%) of all persons who met at least one criterion for polypharmacy had been diagnosed with a mental health disorder at any point in time before the end of 2016. The conditional probability of having a prior mental health diagnosis and meeting criteria for group B was very high (3,196 of 3,727 = 85.8%), compared with that of group A (35.1%) and group C (22.3%).

Covariate analysis showed that, on 30 June 2016, polypharmacy was most preva-lent among active component service mem-bers with the following characteristics: aged 35 years or older (14.9%), female (17.8%), senior enlisted (12.2%) or warrant officer (13.4%) ranks, non-Hispanic black (13.3%), other/unknown marital status (16.3%), an educational attainment of either non-com-pletion of high school (13.1%) or having completed some college (13.8%), 11 or more years of service (13.6%), multiple deploy-ments (12.1%), working in a healthcare occupation (15.3%), and among those diag-nosed with a mental health disorder (20.8%).

Association with suicidal ideation

The frequency distributions of the covariates among cases and controls are shown in Table 2. A total of 2,754 cases of incident SI were identified among ser-vice members with existing mental health

disorders in 2016. A total of 8,262 age- and sex-matched controls were randomly selected from all other service members who had been identified with a diagnosis of a mental health condition. The differences between the distributions of service, rank, race/ethnicity, marital status, educational attainment, length of service, deployment history, and military occupation of cases and controls were statistically significant (p<0.05). In particular, compared to age- and sex-matched controls, cases were more frequently in the Army, junior enlisted (E1–E4), and with less than 2 years of service.

The adjusted odds ratios of polyphar-macy exposure within 1 year prior to date of incident SI diagnosis are presented in Table 3. Among similar service members with existing mental health conditions, those diagnosed with incident SI in 2016 had higher adjusted odds of polypharmacy exposure in the preceding 12 months. The adjusted odds of incident SI were 53% higher for those classified in the “low” stra-tum (AOR 1.53; 95% CI, 1.38–1.71), 120% higher for those in the “moderate” stratum (AOR 2.20; 95% CI, 1.92–2.53), and more than 200% higher for those service mem-bers in the “high” stratum (AOR 3.06; 95% CI, 2.00–4.70).

E D I T O R I A L C O M M E N T

The first part of this report examines the period prevalence of polypharmacy in active component military members dur-ing 2016. The overall period prevalence of 10.8% across the active component mili-tary is higher than the previous estimate of 8.4% from among OEF/OIF veterans in 2011.15 However, it is difficult to compare these two studies that used different defini-tions of polypharmacy. When compared to results from a study using the same defini-tion as this report, the estimated prevalence of 14.5% among Army personnel observed in the current analysis was considerably higher than the previous estimate of 2.2%–7.6% from among active duty Army in combat brigades at Fort Campbell, KY, in 2012.17 One possible explanation for this difference is that the compositions of units with combat-specific missions differ from

June 2018 Vol. 25 No. 6 MSMR Page 5

T A B L E 1 . Number and period prevalence (PP) of polypharmacy, by military and demographic characteristics, active component, U.S. Armed Forces, 2016

1 criterion (low) 2 criteriaa (moderate) 3 criteria (high)

Any group Group A only Group B only Group C only Total Groups A and B

Groups A and C Total Groups A, B,

and CPopulation on 30 June 2016

No. PP (%) No. PP

(%) No. PP (%) No. PP

(%) No. PP (%) No. PP

(%) No. PP (%) No. PP

(%) No. PP (%) No.

Total 139,249 10.8 120,569 9.3 3,727 0.29 197 0.02 124,493 9.6 11,984 0.93 1,796 0.14 13,780 1.07 976 0.08 1,291,250SexMale 102,949 9.5 88,608 8.1 2,853 0.26 156 0.01 91,617 8.4 9,472 0.87 1,178 0.11 10,650 0.98 682 0.06 1,087,226Female 36,300 17.8 31,961 15.7 874 0.43 41 0.02 32,876 16.1 2,512 1.23 618 0.30 3,130 1.53 294 0.14 204,024

Age (years)17–24 45,610 9.2 41,666 8.4 774 0.16 90 0.02 42,530 8.6 1,996 0.40 794 0.16 2,790 0.56 290 0.06 494,20025–34 51,568 10.0 43,946 8.5 1,611 0.31 75 0.01 45,632 8.9 4,863 0.95 641 0.12 5,504 1.07 432 0.08 514,49135+ 42,071 14.9 34,957 12.4 1,342 0.47 32 0.01 36,331 12.9 5,125 1.81 361 0.13 5,486 1.94 254 0.09 282,559

Race/ethnicityNon-Hispanic white 76,937 10.3 65,379 8.8 2,215 0.30 105 0.01 67,699 9.1 7,670 1.03 966 0.13 8,636 1.16 602 0.08 743,669Non-Hispanic black 28,159 13.3 25,061 11.9 631 0.30 42 0.02 25,734 12.2 1,856 0.88 404 0.19 2,260 1.07 165 0.08 211,239Hispanic 19,934 10.5 17,641 9.3 480 0.25 27 0.01 18,148 9.5 1,402 0.74 258 0.14 1,660 0.87 126 0.07 190,167Asian/Pacific Islander 5,309 10.2 4,743 9.1 155 0.30 6 0.01 4,904 9.4 337 0.65 52 0.10 389 0.75 16 0.03 52,000American Indian/Alaska Native 1,342 10.2 1,126 8.5 50 0.38 2 0.02 1,178 8.9 122 0.93 22 0.17 144 1.09 20 0.15 13,185

Other/unknown 7,568 9.3 6,619 8.2 196 0.24 15 0.02 6,830 8.4 597 0.74 94 0.12 691 0.85 47 0.06 80,990EducationLess than high school 297 13.1 241 10.6 13 0.57 1 0.04 255 11.3 35 1.55 5 0.22 40 1.77 2 0.09 2,264High school 85,312 10.5 74,305 9.1 2,187 0.27 149 0.02 76,641 9.4 6,689 0.82 1,293 0.16 7,982 0.98 689 0.08 813,839Some college 22,802 13.8 19,166 11.6 675 0.41 20 0.01 19,861 12.0 2,519 1.52 252 0.15 2,771 1.67 170 0.10 165,532Bachelor's or advanced degree 28,756 10.3 25,019 8.9 796 0.28 23 0.01 25,838 9.2 2,588 0.92 219 0.08 2,807 1.00 111 0.04 280,306

Unknown 2,082 7.1 1,838 6.3 56 0.19 4 0.01 1,898 6.5 153 0.52 27 0.09 180 0.61 4 0.01 29,309Marital statusSingle, never married 46,174 8.5 41,880 7.7 990 0.18 86 0.02 42,956 7.9 2,283 0.42 663 0.12 2,946 0.54 272 0.05 542,195Married 84,536 12.1 71,624 10.3 2,434 0.35 107 0.02 74,165 10.6 8,686 1.25 1,038 0.15 9,724 1.40 647 0.09 696,613Other/unknown 8,539 16.3 7,065 13.5 303 0.58 4 0.01 7,372 14.1 1,015 1.94 95 0.18 1,110 2.12 57 0.11 52,442

ServiceArmy 68,089 14.5 58,295 12.4 2,013 0.43 68 0.01 60,376 12.8 6,356 1.35 826 0.18 7,182 1.53 531 0.11 470,281Navy 25,292 7.8 21,968 6.7 587 0.18 68 0.02 22,623 6.9 1,963 0.60 492 0.15 2,455 0.75 214 0.07 326,073Air Force 32,263 10.4 28,669 9.2 632 0.20 33 0.01 29,334 9.4 2,448 0.79 325 0.10 2,773 0.89 156 0.05 311,527Marine Corps 13,605 7.4 11,637 6.3 495 0.27 28 0.02 12,160 6.6 1,217 0.66 153 0.08 1,370 0.75 75 0.04 183,369

Military rank/gradeJunior enlisted (E1–E4) 58,052 10.4 52,178 9.3 1,215 0.22 103 0.02 53,496 9.6 3,141 0.56 988 0.18 4,129 0.74 427 0.08 559,790Senior enlisted (E5–E9) 60,843 12.2 50,613 10.1 1,948 0.39 77 0.02 52,638 10.5 7,056 1.41 672 0.13 7,728 1.54 477 0.10 500,443Warrant officer (W1–W5) 2,493 13.4 2,122 11.4 75 0.40 1 0.01 2,198 11.8 274 1.48 12 0.06 286 1.54 9 0.05 18,553Junior officer (O1–O3) 9,098 7.0 7,962 6.1 251 0.19 9 0.01 8,222 6.3 759 0.58 76 0.06 835 0.64 41 0.03 129,877Senior officer (O4–O10) 8,763 10.6 7,694 9.3 238 0.29 7 0.01 7,939 9.6 754 0.91 48 0.06 802 0.97 22 0.03 82,587

Length of service (years)Less than 2 25,934 9.5 24,267 8.9 285 0.10 45 0.02 24,597 9.0 803 0.29 407 0.15 1,210 0.44 127 0.05 273,5042–4 33,086 9.4 29,227 8.3 797 0.23 63 0.02 30,087 8.6 2,135 0.61 579 0.17 2,714 0.77 285 0.08 350,3185–10 31,433 10.2 26,417 8.6 1,086 0.35 47 0.02 27,550 9.0 3,235 1.05 389 0.13 3,624 1.18 259 0.08 307,33411 or more 48,796 13.6 40,658 11.3 1,559 0.43 42 0.01 42,259 11.7 5,811 1.61 421 0.12 6,232 1.73 305 0.08 360,094

Deployment historyNever deployed 65,924 10.1 59,138 9.1 1,342 0.21 101 0.02 60,581 9.3 3,850 0.59 1,019 0.16 4,869 0.75 474 0.07 652,8821 deployment 27,796 10.6 23,639 9.0 839 0.32 44 0.02 24,522 9.4 2,729 1.04 338 0.13 3,067 1.17 207 0.08 262,1812 or more deployments 45,529 12.1 37,792 10.0 1,546 0.41 52 0.01 39,390 10.5 5,405 1.44 439 0.12 5,844 1.55 295 0.08 376,187

Military occupationInfantry/artillery/combat engineering/armor 17,880 9.9 15,194 8.4 558 0.31 25 0.01 15,777 8.7 1,773 0.98 192 0.11 1,965 1.09 138 0.08 180,670

Motor transport 4,281 11.5 3,693 9.9 104 0.28 7 0.02 3,804 10.2 365 0.98 79 0.21 444 1.19 33 0.09 37,294Pilot/air crew 2,917 6.0 2,554 5.2 77 0.16 4 0.01 2,635 5.4 259 0.53 9 0.02 268 0.55 14 0.03 48,799Repair/engineer 38,111 9.9 33,107 8.6 878 0.23 67 0.02 34,052 8.8 3,224 0.84 547 0.14 3,771 0.98 288 0.07 385,307Communications/ intelligence 34,976 12.4 30,349 10.7 961 0.34 35 0.01 31,345 11.1 2,978 1.05 421 0.15 3,399 1.20 232 0.08 282,852

Health care 17,665 15.3 15,097 13.1 590 0.51 17 0.01 15,704 13.6 1,609 1.40 216 0.19 1,825 1.58 136 0.12 115,294Other/unknown 23,419 9.7 20,575 8.5 559 0.23 42 0.02 21,176 8.8 1,776 0.74 332 0.14 2,108 0.87 135 0.06 241,034

Existing psychiatric diagnosisb

Yes 54,860 20.8 42,352 16.1 3,196 1.21 44 0.02 45,592 17.3 7,803 2.96 784 0.30 8,587 3.26 681 0.26 263,576No 84,389 8.2 78,217 7.6 531 0.05 153 0.01 78,901 7.7 4,181 0.41 1,012 0.10 5,193 0.51 295 0.03 1,027,674

aNo individuals met criteria for both groups B and C in 2016.bAdjustment disorder, alcohol-related disorder, anxiety disorder, bipolar disorders, depressive disorder, personality disorders, psychoses, post-traumatic stress disorder, schizophrenia, and/or substance-related disorder on or before 31 December 2016

MSMR Vol. 25 No. 6 June 2018 Page 6

that of the Army overall and, over time, select for healthier soldiers who do not have deployment-limiting conditions. Soldiers with complex medical problems that limit deployment, who are more likely to require multiple medications and potentially meet criteria for polypharmacy, are less likely to serve in such a unit and thus the prevalence estimate would be lower. Furthermore, it is possible that the prevalence of prescription drug use and polypharmacy have increased since 2012.

The observation that individuals in group B (those with multiple psychotro-pic or CNSD drugs) had the highest con-ditional probability of a comorbid mental health disorder is expected. Individuals who are prescribed multiple psychotropic drugs are more likely to have one or more mental health conditions for which these types of medications are clinically indi-cated. Similarly, the fact that most (86.6%) of the polypharmacy exposure in this study arose from individuals meeting criteria for group A is not surprising as there are many more types of clinical situations for which sufficient medications to satisfy the criteria for group A would be prescribed than there are for the other two groups.

The observation that higher preva-lence of polypharmacy exposure was found among individuals working in health-care occupations echoes the observation that healthcare workers accounted for the highest rates of overall prescription drugs in 2014.51 These findings highlight a trend seen in surveillance reports on low back pain,52 acute respiratory illness,53 alcohol-related diagnoses,54 and obstructive sleep apnea,55 among others. A possible explana-tion is that healthcare workers have easier access to medical treatment by virtue of working in the same facility. Alternatively, persons with medical needs may select for healthcare occupations that are generally less physically demanding than combat-specific or other support occupations.

The second part of this report explores the association between incident SI and precedent polypharmacy exposure within 12 months. Polypharmacy was found to be associated with incident SI (AOR = 1.53–3.06), which supports similar findings from a 2010–2011 study of OEF/OIF veterans.15 The strength of the association increased as

T A B L E 2 . Comparison of cases of suicidal ideation and selected controls, by military and demographic characteristics, service members with a previous diagnosis of mental health disorder, active component, U.S. Armed Forces, 2016

Case (n=2,754) Control (n=8,262)

No. % No. %SexMale 2,007 72.9 6,021 72.9Female 747 27.1 2,241 27.1

Age (years)17–24 1,339 48.6 4,017 48.625–34 1,003 36.4 2,996 36.335+ 412 15.0 1,249 15.1

Race/ethnicityNon-Hispanic white 1,494 54.3 4,686 56.7Non-Hispanic black 589 21.4 1,581 19.1Hispanic 387 14.1 1,203 14.6Asian/Pacific Islander 127 4.6 258 3.1American Indian/Alaska Native 22 0.8 94 1.1Other/unknown 135 4.9 440 5.3

Marital statusSingle, never married 1,201 43.6 3,337 40.4Married 1,364 49.5 4,404 53.3Other/unknown 189 6.9 521 6.3

EducationLess than high school 6 0.2 14 0.2High school 2,129 77.3 6,150 74.4Some college 364 13.2 1,109 13.4Bachelor's or advanced degree 222 8.1 891 10.8Unknown 33 1.2 98 1.2

ServiceArmy 1,639 59.5 4,123 49.9Navy 350 12.7 1,587 19.2Air Force 517 18.8 1,616 19.6Marine Corps 248 9.0 936 11.3

GradeJunior enlisted (E1–E4) 1,695 61.6 4,446 53.8Senior enlisted (E5–E9) 939 34.1 3,220 39.0Warrant officer (W1–W5) 8 0.3 78 0.9Junior officer (O1–O3) 83 3.0 313 3.8Senior officer (O4–O10) 29 1.1 205 2.5

Length of service (years)Less than 2 650 23.6 1,320 16.02–4 914 33.2 2,969 35.95–10 680 24.7 2,090 25.311 or more 510 18.5 1,883 22.8

Deployment historyNever deployed 1,683 61.1 4,764 57.71 deployment 474 17.2 1,539 18.62 or more deployments 597 21.7 1,959 23.7

Military occupationInfantry/artillery/combat engineering/armor 412 15.0 1,127 13.6

Motor transport 102 3.7 326 4.0Pilot/air crew 22 0.8 64 0.8Repair/engineering 751 27.3 2,266 27.4Communications/intelligence 654 23.8 2,091 25.3Health care 352 12.8 1,121 13.6Other/unknown 461 16.7 1,267 15.3

Polypharmacy exposure1 criterion (low) 814 29.6 2,002 24.22 criteria (moderate) 442 16.1 736 8.93 criteria (high) 46 1.7 49 0.6

June 2018 Vol. 25 No. 6 MSMR Page 7

more criteria for polypharmacy exposure were satisfied, although the CIs widen to reflect the smaller sample sizes. This associ-ation should be interpreted with some cau-tion as there are limitations to this report. It is possible that this observation reflects the influence of confounding from some variable that was not included in this anal-ysis such as financial stress, recent loss of a loved one, baseline variations in accession standards by year of entry into the mili-tary, or perhaps most importantly, the type or severity of the underlying mental health disorder. For example, individuals with more severe mental health disorders might be more likely to report SI to a healthcare provider and also more likely to be pre-scribed more medications, which would confound association between polyphar-macy and incident SI found in this report. The administrative data used for this analy-sis have limited capacity to ascertain illness severity and future study should take this into consideration.

Another limitation of the current anal-ysis is that it did not account for differences between individuals who meet polyphar-macy criteria repeatedly (e.g., chronic pre-scription use) compared with those who only meet criteria transiently (e.g., follow-ing an acute injury or surgical procedure). The duration of polypharmacy exposure may serve as an indicator of conditions that are more chronic and that previous research has demonstrated increase the risk of suicidal behaviors.56,57 Therefore, poly-pharmacy may have been misclassified in this analysis, which could have biased the observed association with SI toward null. Future studies should explore the difference between chronic and acute polypharmacy on outcomes such as SI. Finally, this report did not capture medications dispensed in the deployed environment or aboard ship for the U.S. Navy, so the true period preva-lence may have been underestimated.

This report describes the findings of an exploratory analysis of the prevalence of polypharmacy among active component service members of the U.S. Armed Forces. The observation of an independent asso-ciation between incident SI and precedent polypharmacy in the previous year is con-cerning. However, there are important lim-itations to the current analysis that should

T A B L E 3 . Crude (OR) and adjusteda odds ratio (AOR) of suicidal ideation, active compo-nent service members with a previous diagnosis of mental health disorder, U.S. Armed Forces, 2016

OR 95% CI AOR 95% CI

Polypharmacy exposure1 criterion (low) 1.32 (1.20–1.45) 1.53 (1.38–1.71)2 criteria (moderate) 1.97 (1.73–2.24) 2.20 (1.92–2.53)3 criteria (high) 2.84 (1.90–4.26) 3.06 (2.00–4.70)

Race/ethnicityNon-Hispanic white ref refNon-Hispanic black 1.18 (1.05–1.32) 1.08 (0.95–1.21)Hispanic 1.01 (0.89–1.15) 1.00 (0.87–1.14)Asian/Pacific Islander 1.55 (1.24–1.93) 1.46 (1.16–1.85)American Indian/Alaska Native 0.73 (0.46–1.16) 0.81 (0.49–1.32)Other/unknown 0.96 (0.79–1.18) 1.13 (0.91–1.40)

Marital statusSingle, never married ref refMarried 0.82 (0.73–0.91) 0.84 (0.75–0.94)Other/unknown 0.95 (0.78–1.16) 0.95 (0.77–1.18)

EducationLess than high school 1.15 (0.44–3.00) 1.43 (0.52–3.94)High school ref refSome college 0.88 (0.77–1.02) 0.87 (0.75–1.01)Bachelor's or advanced degree 0.65 (0.55–0.78) 0.65 (0.52–0.83)Unknown 0.94 (0.63–1.41) 1.06 0.69–1.63)

ServiceArmy ref refNavy 0.55 (0.48–0.63) 0.58 (0.50–0.67)Air Force 0.80 (0.71–0.90) 0.88 (0.77–1.00)Marine Corps 0.67 (0.57–0.78) 0.76 (0.64–0.89)

Military rank/gradeJunior enlisted (E1–E4) ref refSenior enlisted (E5–E9) 0.45 (0.39–0.52) 0.65 (0.54–0.77)Warrant officer (W1–W5) 0.13 (0.06–0.28) 0.20 (0.09–0.43)Junior officer (O1–O3) 0.42 (0.32–0.55) 0.61 (0.43–0.86)Senior officer (O4–O10) 0.16 (0.10–0.25) 0.33 (0.20–0.54)

Length of service (years)Less than 2 ref ref2–4 0.42 (0.36–0.49) 0.42 (0.36–0.49)5–10 0.28 (0.23–0.34) 0.35 (0.28–0.45)11 or more 0.14 (0.11–0.18) 0.19 (0.14–0.26)

Deployment historyNever deployed ref ref1 deployment 0.79 (0.69–0.90) 0.99 (0.85–1.15)2 or more deployments 0.73 (0.63–0.85) 1.01 (0.84–1.21)

Military occupationInfantry/artillery/combat engineering/armor 1.01 (0.86–1.18) 0.94 (0.80–1.12)

Motor transport 0.86 (0.67–1.10) 0.79 (0.61–1.03)Pilot/air crew 0.94 (0.57–1.56) 1.78 (1.03–3.05)Repair/engineering 0.91 (0.80–1.05) 0.96 (0.83–1.11)Communications/intelligence 0.86 (0.75–0.98) 0.86 (0.74–0.99)Health care 0.86 (0.73–1.01) 0.88 (0.74–1.04)Other/unknown ref ref

CI, confidence intervalaConditional logistic regression of age- and sex-matched groups, adjusted for listed covariates with indicated reference groups

MSMR Vol. 25 No. 6 June 2018 Page 8

be addressed in future studies before trying to infer a causal relationship. Nevertheless, it would be prudent to screen for SI during the polypharmacy clinical encounter, par-ticularly for persons with any of the mental health disorders considered in this report. Inclusion of DoDSER data along with med-ically coded SI in future surveillance would serve to increase the sensitivity of identify-ing incident cases of SI. Author affiliations: Walter Reed Army Insti-tute of Research, Silver Spring, MD (MAJ Eide); Armed Forces Health Surveillance Branch, Defense Health Agency, Silver Spring, MD (Dr. Stahlman).

Acknowledgments: The authors thank Dr. Saixia Ying for her invaluable assistance with data collection and analysis; COL (Dr.) P. Ann Loveless for project oversight, mentorship, and feedback; and Ms. Amaris Thurston for project coordination and main-taining timeline.

Disclaimer: Material has been reviewed by the Walter Reed Army Institute of Research. There is no objection to its presentation and/or publication. The opinions or assertions contained herein are the private views of the author(s), and are not to be construed as official, or as reflecting the true views of the Department of the Army or the Department of Defense. The investigators have adhered to the policies for protection of human subjects as prescribed in AR 70-25.

R E F E R E N C E S

1. Gillette C, Prunty  L, Wolcott  J,  Broedel-Zaugg  K. A new lexicon for polypharmacy: implications for research, practice, and education. Res Social Adm Pharm. 2015;11(3):468–471.2. Kouladjian L, Hilmer SN, Chen TF, Le Couteur DG, Gnjidic D. Assessing the harms of polypharma-cy requires careful interpretation and consistent defi-nitions. Br J Clin Pharmacol. 2014;78(3):670–671.3. Kantor ED, Rehm CD, Haas JS, Chan AT, Giovannucci EL. Trends in prescription drug use among adults in the United States from 1999–2012. JAMA. 2015;314(17):1818–1831.4. Alhawassi TM, Krass I, Bajorek BV, Pont LG. A systematic review of the prevalence and risk factors for adverse drug reactions in the elderly in the acute care setting. Clin Interv Aging. 2014;9:2079–2086.5. Hajjar ER, Cafiero AC, Hanlon JT. Polypharma-cy in elderly patients. Am J Geriatr Pharmacother. 2007;5(4):345–351.

6. Linton A, Garber M, Fagan NK, Peterson MR. Examination of multiple medication use among TRICARE beneficiaries aged 65 years and older. J Manag Care Pharm. 2007;13(2):155–162.7. Maher RL, Hanlon J, Hajjar ER. Clinical con-sequences of polypharmacy in elderly. Expert Opin Drug Saf. 2014;13(1):57–65.8. Mannucci PM, Nobili A. Multimorbidity and polypharmacy in the elderly: lessons from Reposi. Intern Emerg Med. 2014;9(7):723–734.9. Fulton MM, Allen ER. Polypharmacy in the el-derly: a literature review. J Am Acad Nurse Pract. 2005;17(4):123–132.10. Rollason V, Vogt N. Reduction of polypharmacy in the elderly: a systematic review of the role of the pharmacist. Drugs Aging. 2003;20(11):817–832.11. 2016 Demographics Report: Profile of the Mili-tary Community. In: Office of the Deputy Assistant Secretary of Defense for Military Community and Family Policy (OADASD (MC&FP)), ed. Washing-ton, DC: Department of Defense; 2016.12. Gallaway MS, Lagana-Riordan C, Dabbs CR, et al. A mixed methods epidemiological investiga-tion of preventable deaths among U.S. Army sol-diers assigned to a rehabilitative warrior transition unit. Work. 2015;50(1):21–36.13. Lande RG, Gragnani C. Relationships between polypharmacy and the sleep cycle among active-duty service members. J Am Osteopath Assoc. 2015;115(6):370–375.14. Golub A, Bennett AS. Prescription opioid ini-tiation, correlates, and consequences among a sample of OEF/OIF military personnel. Subst Use Misuse. 2013;48(10):811–820.15. Collett GA, Song K, Jaramillo CA, Potter JS, Finley EP, Pugh MJ. Prevalence of central ner-vous system polypharmacy and associations with overdose and suicide-related behaviors in Iraq and Afghanistan war veterans in VA care 2010–2011. Drugs Real World Outcomes. 2016;3(1):45–52.16. Defense Health Agency. Polypharmacy Medi-cation Analysis and Report Tool (Poly-Mart). 2017; https://health.mil/PolyMART. Accessed on 19 Octo-ber 2017.17. Ridderhoff KJ, Hull JR, Sandberg SK. Blanch-field Army Community Hospital Polypharmacy Clin-ic. J Manag Care Spec Pharm. 2015;21(1):8–11.18. Office of the Surgeon General/Medical Com-mand. Policy Memo 15-039: Guidance for Man-aging Polypharmacy and Preventing Medication Overdose in Patients Prescribed Psychotropic and Central Nervous System Depressants. 2015.19. Giummarra MJ, Gibson SJ, Allen AR, Pichler AS, Arnold CA. Polypharmacy and chronic pain: harm exposure is not all about the opioids. Pain Med. 2015;16(3):472–479.20. Corsonello A, Abbatecola AM, Fusco S, et al. The impact of drug interactions and polypharmacy on antimicrobial therapy in the elderly. Clin Micro-biol Infect. 2015;21(1):20–26.21. Guthrie B, Makubate B, Hernandez-Santiago V, Dreischulte T. The rising tide of polypharmacy and drug-drug interactions: population database analysis 1995–2010. BMC Med. 2015;13:74.22. Kotlinska-Lemieszek A, Klepstad P, Haugen DF. Clinically significant drug-drug interactions in-volving opioid analgesics used for pain treatment in patients with cancer: a systematic review. Drug Des Devel Ther. 2015;9:5255–5267.23. Barnes TR, Paton C. Antipsychotic polyphar-macy in schizophrenia: benefits and risks. CNS Drugs. 2011;25(5):383–399.

24. Calderon-Larranaga A, Gimeno-Feliu LA, Gon-zalez-Rubio F, et al. Polypharmacy patterns: unrav-elling systematic associations between prescribed medications. PLoS One. 2013;8(12):e84967.25. Correll CU, Detraux J, De Lepeleire J, De Hert M. Effects of antipsychotics, antidepressants and mood stabilizers on risk for physical diseases in people with schizophrenia, depression and bipolar disorder. World Psychiatry. 2015;14(2):119–136.26. Madan A, Oldham JM, Gonzalez S, Fowler JC. Reducing adverse polypharmacy in patients with borderline personality disorder: an empirical case study. Prim Care Companion CNS Disord. 2015;17(4).27. Everett WW. Polypharmacy and adverse drug-related events. Ann Emerg Med. 2003;41(2): 278–279.28. Hein C, Forgues A, Piau A, Sommet A, Vellas B, Nourhashemi F. Impact of polypharmacy on oc-currence of delirium in elderly emergency patients. J Am Med Dir Assoc. 2014;15(11):850.e811–855.29. Hohl CM, Dankoff J, Colacone A, Afilalo M. Polypharmacy, adverse drug-related events, and potential adverse drug interactions in elderly pa-tients presenting to an emergency department. Ann Emerg Med. 2001;38(6):666–671.30. Makris UE, Pugh MJ, Alvarez CA, et al. Ex-posure to high-risk medications is associated with worse outcomes in older veterans with chronic pain. Am J Med Sci. 2015;350(4):279–285.31. Samaras N, Chevalley T, Samaras D, Gold G. Older Patients in the emergency department: a re-view. Ann Emerg Med. 2010;56(3):261–269.32. Xu J, Murphy SL, Kochanek KD, Arias E. Morl-ity in the United States, 2015. NCHS Data Brief. 2016;(267):1–8.33. Curtin SC, Warner M, Hedegaard H. Increase in suicide in the United States, 1999–2014. NCHS Data Brief. 2016;(241):1–8.34. Ramchand R, Acosta J, Burns RM, Jaycox LH, Pernin CG. The war within: preventing suicide in the U.S. military. Rand Health Q. 2011;1(1):2.35. Lineberry TW, O'Connor SS. Suicide in the U.S. Army. Mayo Clin Proc. 2012;87(9):871–878.36. Defense Suicide Prevention Office. CY 2015 Department of Defense Suicide Event Report (DoDSER) Annual Report. Washington, DC. 2016.37. National Center for Telehealth and Technol-ogy. Department of Defense Suicide Event Report. 2017; https://t2health.dcoe.mil/programs/dodser. Accessed on 26 October 2017.38. Ursano RJ, Colpe LJ, Heeringa SG, Kessler RC, Schoenbaum M, Stein MB. The Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS). Psychiatry. 2014;77(2):107–119.39. Defense Suicide Prevention Office. History of the Defense Suicide Prevention Office. 2017; http://www.dspo.mil/About-DSPO/History/. Accessed on 26 October 2017.40. LeardMann CA, Powell TM, Smith TC, et al. Risk factors associated with suicide in cur-rent and former U.S. military personnel. JAMA. 2013;310(5):496–506.41. Schoenbaum M, Kessler RC, Gilman SE, et al. Predictors of suicide and accident death in the Army Study to Assess Risk and Resilience in Ser-vicemembers (Army STARRS): Results from the Army Study to Assess Risk and Resilience in Ser-vicemembers (Army STARRS). JAMA Psychiatry. 2014;71(5):493–503.42. Reger MA, Smolenski DJ, Skopp NA, et al. Risk of suicide among U.S. military service mem-

June 2018 Vol. 25 No. 6 MSMR Page 9

CE/CME

This activity offers continuing education (CE) and continuing medical education (CME) credit to qualified professionals, as well as a certificate of participation to those desiring documentation. For more information, go to www.health.mil/msmrce.

Key points• The highest levels of polypharmacy were among Army members and the lowest levels were among those in the Navy or the

Marine Corps.

• On 30 June 2016, polypharmacy was most prevalent among active component service members with the following characteris-tics: aged 35 years or older, female, senior enlisted or warrant officer ranks, non-Hispanic black, other/unknown marital status, an educational attainment of either non-completion of high school or having completed some college, 11 or more years of service, multiple deployments, working in a healthcare occupation, and among those diagnosed with a mental health disorder.

• Among similar service members with existing mental health conditions, those diagnosed with incident suicide ideation (SI) in 2016 had higher adjusted odds of polypharmacy exposure in the preceding 12 months. The adjusted odds of incident SI were 53% higher for those classified in the “low” polypharmacy stratum (met one of the following criteria: had four or more prescriptions all within 30 days of one another for any drug [including antibiotics, antihypertensives, statins, etc.] where at least one was an opioid; or had four or more prescriptions all within 30 days of one another from among the seven categories of psychotropic or CNSD medications; or had three or more emergency room visits during the surveillance period, each within 7 days of an opioid prescription). The adjusted odds of incident SI more than doubled for service members in the “moderate” polypharmacy stratum (met two of the three criteria) and tripled among those in the “high” polypharmacy stratum (met all three criteria).

Learning objectives

1. The reader will understand the period prevalence of polypharmacy involving opioid, CNS depressant, and psychotropic medica-tions among active component service members in the U.S. military in 2016.

2. The reader will interpret the distribution of polypharmacy involving opioid, CNS depressant, or psychotropic medications by various demographic and service-related variables.

3. The reader will explain the association between polypharmacy and incident suicidal ideation among service members with men-tal health disorders.

Disclosures: MSMR staff authors, DHA J7, AffinityCE/PESG, as well as the planners and reviewers of this activity have no financial or non-financial interest to disclose.

bers following Operation Enduring Freedom or Operation Iraqi Freedom Deployment and sepa-ration from the U.S. military. JAMA Psychiatry. 2015;72(6):561–569.43. Nock MK, Stein MB, Heeringa SG, et al. Prev-alence and correlates of suicidal behavior among soldiers: results from the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS). JAMA Psychiatry. 2014;71(5):514–522.44. Ursano RJ, Heeringa SG, Stein MB, et al. Prevalence and correlates of suicidal behavior among new soldiers in the U.S. Army: results from the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS). Depress Anxi-ety. 2015;32(1):3–12.45. Ursano RJ, Kessler RC, Heeringa SG, et al. Nonfatal suicidal nehaviors in U.S. Army ad-ministrative records, 2004–2009: Results from the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS). Psychiatry. 2015;78(1):1–21.46. Kessler RC, Borges G, Walters EE. Prevalence of and risk factors for lifetime suicide attempts in the National Comorbidity Survey. Arch Gen Psychiatry. 1999;56(7):617–626.

47. Ursano RJ, Kessler RC, Stein MB, et al. Medically documented suicide ideation among U.S. Army Soldiers. Suicide Life Threat Behav. 2017;47(5):612–628.48. Nweke N, Spiess A, Corrigan E, et al. Sur-veillance of Suicidal Behavior, January through December 2015. Aberdeen Proving Ground, MD: Army Public Health Center; 2016. https://phc.amedd.army.mil/PHC%20Resource%20Library/2015SurveillanceofSuicidalBehavior.pdf.49. U.S. Food and Drug Administration. Structured Product Labeling: Pharmacologic Class. 2015; https://www.fda.gov/ForIndustry/DataStandards/StructuredProductLabeling/ucm162549.htm. Ac-cessed on 20 October 2017.50. Armed Forces Health Surveillance Branch. Surveillance Case Definitions. 2016; https://www.health.mil/Military-Health-Topics/Health-Readi-ness/Armed-Forces-Health-Surveillance-Branch/Epidemiology-and-Analysis/Surveillance-Case-Definitions. Accessed on 20 October 2017.51. Hurt L, Zhong X. Brief report: Rate of pre-scriptions by therapeutic classification, active component, U.S. Armed Forces, 2014. MSMR. 2015;22(9):17–21.

52. Clark LL, Hu Z. Diagnoses of low back pain, ac-tive component, U.S. Armed Forces, 2010–2014. MSMR. 2015;22(12):8–11.53. Brundage JF, Taubman SB, Clark LL. Rates of acute respiratory illnesses of infectious and al-lergic etiologies after permanent changes of duty assignments, active component, U.S. Army, Air Force, and Marine Corps, January 2005–Septem-ber 2015. MSMR. 2015;22(11):2–7.54. Armed Forces Health Surveillance Cen-ter. Alcohol-related diagnoses, active compo-nent, U.S. Armed Forces, 2001–2010. MSMR. 2011;18(10):9–13.55. Rogers AE, Stahlman S, Hunt DJ, Oh GT, Clark LL. Obstructive sleep apnea and associated attri-tion, active component, U.S. Armed Forces, Janu-ary 2004–May 2016. MSMR. 2016;23(10):2–11.56. Hooley JM, Franklin JC, Nock MK. Chronic pain and suicide: understanding the association. Curr Pain Headache Rep. 2014;18(8):435.57. Karasouli E, Latchford G, Owens D. The im-pact of chronic illness in suicidality: a qualitative exploration. Health Psychol Behav Med. 2014;2(1): 899–908.

MSMR Vol. 25 No. 6 June 2018 Page 10

This report uses routinely collected data in the Defense Medical Surveillance System (DMSS) to characterize the incidence and burden of medical condi-tions throughout the careers of service members separating from the active component of the U.S. Armed Forces between 1 October 2014 and 30 Sep-tember 2015. Three surveillance periods between 30 September 2000 and 30 September 2015 were defined by each individual’s time in service: early (first 6 months), middle (middle 6 months), and late (last 6 months). Overall, sep-arating service members were most often aged 25–34 years (59.4%), male (84.0%), non-Hispanic white (64.0%), junior enlisted (52.4%), in the Marine Corps (33.1%), serving in a repair/engineering occupation (33.0%), and had never deployed (52.5%). The top five burden of disease categories across surveillance periods by sex were very similar, including mental health dis-orders, which exhibited similar upward trends across the three surveillance periods (males: 1.3%, 17.0%, and 35.6%; females: 1.8%, 15.1%, and 32.4%, respectively). The most common diagnoses exhibiting upward, downward, or bimodal trends by incidence rate differences were mental health disorders, respiratory infections/diseases, and musculoskeletal diseases, respectively.

Variations in the Incidence and Burden of Illnesses and Injuries Among Non-retiree Service Members in the Earliest, Middle, and Last 6 Months of Their Careers, Active Component, U.S. Armed Forces, 2000–2015Colby C. Uptegraft, MD (Capt, USAF); Shauna Stahlman, PhD, MPH

The burden and incidence of medical conditions among residents of the U.S. vary by age, sex, race/ethnic-

ity, socioeconomic status, insurance status, occupation, clinical setting, geographical region, and time.1-7 Active component ser-vice members (ACSMs) represent a unique subset of this population. Entrance into the U.S. Armed Forces requires a minimum level of health and fitness among appli-cants,8 making ACSMs, at least upon entry, healthier than their civilian counterparts, a phenomenon known as the “healthy sol-dier effect.”9,10 After entry, occupational and readiness requirements expose ser-vice members to numerous unique hazards not commonly found outside the military. The selection of healthy men and women for accession and the subsequent exposure

profile of active service prevent the gener-alization of civilian incidence and burden findings to military populations.

Military medical standards for acces-sion are stricter than retention standards. For instance, potential recruits with symp-tomatic hemorrhoids and/or abdominal wall hernias would likely not meet acces-sion standards, but ACSMs with these con-ditions, with rare exceptions, would not be medically separated.8,11-13 However, ACSMs must continue to meet health and fitness requirements to maintain their service eli-gibility. These requirements are updated periodically and vary across occupational class and branch of service.8,11-13 The dis-tinct occupations within each service expose ACSMs to unique work settings and environments, both in garrison and

while deployed. For those ACSMs pursu-ing career longevity and certain military occupations, there are disincentives for reporting medical conditions or seeking care. The development of post-accession medical conditions raises the possibility of adverse personnel actions such as medical evaluation boards and subsequent medi-cal separation, duty location or deployment limitations, or career field denial or termi-nation, particularly special-duty occupa-tions such as aviation and special forces.

On the other hand, as ACSMs approach and prepare for retirement (ser-vice of 20 or more years) or separation (ser-vice of less than 20 years), there are positive incentives for reporting health issues and having them evaluated, treated if necessary, and documented in their health records. Service members who retire or separate with service-connected health conditions or disabilities are eligible for disability com-pensation and health care for those condi-tions through the Department of Veterans Affairs.14 Because of the occupational dis-incentives for healthcare-seeking behaviors and the pre-separation incentives for such behaviors, precise estimates of the bur-den and incidence of medical conditions throughout service members’ careers are difficult to ascertain.

Two previous studies characterized the incidence of illnesses and injuries immedi-ately prior to retirement. Service members within 6 months of retirement were more likely than pre-retirees (12–18 months before retirement) to receive any medi-cal diagnosis, particularly those diagnoses common among similarly aged Americans and compensable as service-connected dis-abilities.15 Additionally, 72.1% of retirees were diagnosed with a new medical condi-tion within 6 months of retirement, and the number of illnesses and injuries, both new and old, differed by occupational group and

June 2018 Vol. 25 No. 6 MSMR Page 11

rank.16 However, these studies included only retirees, and, on average, about 17% of enlisted service members and officers across all branches ever reach retirement eligibility.17 The remaining 83% of ACSMs that separate from service tend to be younger, lower ranking, and less advanced in their careers than retirees. Although other differences likely exist, these alone make separating service members a distinct subpopulation, warranting further study to better characterize illness and injury across the careers of ACSMs.

Medical readiness is the core focus of the Military Health System.18,19 Service-specific databases that log readiness sta-tistics indicate that deployment-limiting medical conditions are the main reasons why ACSMs are not deployable and thus not medically ready.20-22 Characterization of the temporality of when medical con-ditions occur throughout the career of ACSMs, stratified by demographic and military characteristics, may offer insight for preventive interventions to improve the health and medical readiness of service populations. The purpose of this study is to characterize the incidence and burden of medical conditions throughout the careers of separating service members.

M E T H O D S

The surveillance population included all individuals who entered active military service after 30 September 2000 and sep-arated from the active component of the Air Force, Army, Marine Corps, or Navy between 1 October 2014 and 30 Septem-ber 2015, and who had at least 48 months of continuous active service at the time of separation. Three surveillance periods were determined by the individual’s dates of ser-vice: early (first 6 months), middle (middle 6 months), and late (last 6 months of ser-vice). Retirees (service members with 240 or more months of active service) and ser-vice members with any breaks in service during their careers (n = 9,585) or deploy-ment days during any of the three surveil-lance periods (n = 16,339) were excluded.

For each individual in the study pop-ulation, all illness and injury diagnoses

(ICD-9: 000–999) that were recorded during inpatient and outpatient medi-cal encounters in U.S. military medical facilities, and from purchased-care pro-viders, were obtained from standardized medical records routinely maintained in the Defense Medical Surveillance System (DMSS). All ICD codes were grouped into 25 major burden of disease categories based on a modified version23 of the Global Bur-den of Disease Study.24 This grouping was done to provide an overall estimate of the most common conditions affecting service men and women during the three different surveillance periods.

To calculate the total hospitalizations and outpatient visits for each diagnosis, only three-digit ICD-9 codes in the first diagnostic position were included, with no more than one encounter per three-digit ICD code per individual per day. The pro-portion of total encounters for each burden of disease category was calculated by divid-ing the total encounters for each category by the overall total number of encounters.

Three-digit ICD-9 codes documented in any diagnostic position recorded for the first time in an ACSM’s career and during each surveillance period were considered incident (first-time) occurrences. The total number of at-risk ACSMs for each surveil-lance period was the number of service members in the study population not diag-nosed with the respective ICD code prior to the start of the surveillance period. Inci-dence rates (IRs) were calculated by divid-ing the number of first-time occurrences by the number of at-risk service members for each ICD code for each surveillance period.

Three trends in IR—upward, down-ward, and bimodal—were chosen for anal-yses. An upward trend was defined as an increase in IR between both the early and middle and middle and late surveillance periods. Diagnoses exhibiting an upward trend were ranked by the total difference in the IR between the late and early surveil-lance periods from largest to smallest. A downward trend was defined as a decrease in the IR between both the early and mid-dle and middle and late surveillance peri-ods. Diagnoses exhibiting a downward trend were ranked by the total difference in the IR between the late and early surveil-lance periods from largest to smallest. A

bimodal trend was defined as a decrease in the IR between the early and middle sur-veillance period and an increase between the middle and late period. Diagnoses exhibiting a bimodal trend in the IRs were ranked by the sum of the absolute value of the difference between the early and middle surveillance period and the absolute value of the difference between the middle and late period from largest to smallest. For all trends, results were stratified by sex.

R E S U L T S

The demographic and military charac-teristics of the study population by time in service are shown in Table 1. Of the 45,363 total separating service members between 1 October 2014 and 30 September 2015 with at least 4 years of continuous active compo-nent service, 32,597 (71.9%) separated with 4–8 years of active service, 10,040 (22.1%) with 8–12 years, and 2,726 (6.0%) with 12–15 years. Overall, separating service members were most often aged 25–34 years (59.4%), male (84.0%), non-Hispanic white (64.0%), and junior enlisted (52.4%). The Marine Corps (33.1%) contributed more ACSMs to the study population than any of the other services. The occupational cat-egory of repair/engineering was the most common (33.0%). Among all members of the study population, 52.5% had never deployed.

Overall, females had more total encounters per service member in the early, middle, and late surveillance peri-ods (5.0, 6.3, and 12.3, respectively) than males (3.0, 2.6, and 7.3, respectively) (data not shown). The proportions of total outpa-tient and inpatient encounters by burden of disease major categories for each surveil-lance period and sex are shown in the Fig-ure. The five top-ranking burden of disease major categories in all surveillance periods in both sexes were very similar. Injuries/poisonings, musculoskeletal diseases, and signs and symptoms were present in the five top-ranking categories in both sexes across all three surveillance periods. Respiratory infections (males: 29.9%; females: 17.0%) were among the top-ranking diagnoses in the first surveillance period but were

MSMR Vol. 25 No. 6 June 2018 Page 12

E D I T O R I A L C O M M E N T

This report examines the diagnos-tic trends across three time points in the careers of service members separating from the active component between 1 October 2014 and 30 September 2015. Several com-parisons can be made to previous studies on overall disease burdens and common diagnoses in recruits and retiring service members.

Crowded living conditions25 and the stressful environment of military train-ing,26-29 make recruit populations particu-larly susceptible to respiratory infections.30 Therefore, the observed encounter bur-den of respiratory infections in the first 6 months of service and the downward trend in related three-digit ICD-9 codes across surveillance periods is consistent with expectations and prior findings. Addition-ally, the commonality of musculoskeletal conditions, including “other and unspeci-fied disorders of joint” (ICD-9: 719), “other disorders of soft tissues” (ICD-9: 729), and “disorders of muscle ligament and fascia” (ICD-9: 728), in both the first 6 months and last 6 months of service (bimodal trend) is not surprising. Two factors are likely con-tributing to this finding. First, in the initial 6 months of service, recruits during basic training are particularly prone to inju-ries and musculoskeletal conditions31-34; this increased risk is especially evident for females,34-36 which might also explain the greater proportionate encounter bur-den for injuries/poisonings and musculo-skeletal disorders among females during the first 6 months (38.4%) versus among males (29.5%). During the last 6 months of service, ACSMs are encouraged to report chronic conditions during separation/retirement physicals as documentation of service-connected conditions increases their likelihood of both obtaining disability compensation14 through the Department of Veterans Affairs (VA) and accessing VA healthcare.37 Conversely, during the mid-dle 6 months of service, other than being older and having more time in service than the first 6 months of service, there are not any strong risk factors or VA entitlement incentives for seeking care or reporting these conditions.

T A B L E 1 . Demographic and military characteristicsa of service members who separated between 1 October 2014 and 30 September 2015 with 4–15 years of continuous active service, active com-ponent, U.S. Armed Forces

No. %Total individuals 45,363 100.0SexMale 38,092 84.0Female 7,271 16.0

Age (years) at time of separation17–24 15,788 34.825–34 26,936 59.435–39 1,934 4.340–49 678 1.550+ 27 0.1

Race/ethnicityNon-Hispanic white 29,028 64.0Non-Hispanic black 5,471 12.1Hispanic 5,868 12.9Asian/Pacific Islander 1,632 3.6American Indian/Alaska Native 606 1.3

Other/unknown 2,758 6.1ServiceArmy 13,414 29.6Navy 10,046 22.1Air Force 6,879 15.2Marine Corps 15,024 33.1

Grade at time of separationJunior enlisted (E1–E4) 23,761 52.4Senior enlisted (E5–E9) 18,424 40.6Warrant officer (W1–W5) 76 0.2Junior officer (O1–O3) 2,500 5.5Senior officer (O4–O10) 602 1.3

Military occupationInfantry/artillery/combat engineering/armor 7,901 17.4

Motor transport 1,444 3.2Pilot/air crew/air traffic 659 1.5Repair/engineering 14,969 33.0Communications/ intelligence 9,876 21.8

Health care 4,365 9.6Other/unknown 6,149 13.6

Time in service (years)4–8 32,597 71.98–12 10,040 22.112–15 2,726 6.0

No. of deploymentsNone 23,828 52.5One 12,689 28.0More than one 8,846 19.5

aCharacteristics at the time of military separation

not among the five top-ranking diagno-ses thereafter. Additionally, mental health disorders exhibited similar upward trends across the three surveillance periods rep-resenting one-third of total encounters in both sexes in the last period (males: 1.3%, 17.0%, and 35.6%; females: 1.8%, 15.1%, and 32.4%, respectively).

Although the distributions of the major burden of disease categories in the surveillance periods were broadly similar for males and females, several differences emerged. Injuries/poisonings remained consistent at 15.5% across the first and middle surveillance period and then fell to 6.6% among males but showed a consistent decline across surveillance periods among females (15.0%, 7.9%, and 4.2%, respec-tively). Musculoskeletal disorders peaked at 24.3% in the middle surveillance period for males but peaked at 23.4% in the first period for females. The “all other” cate-gory (the proportion of total encounters for all major burden categories ranked 6th through 25th for each surveillance period) was also slightly larger among females than among males across all periods (males: 23.2%, 24.8%, and 16.1%; females: 25.4%, 31.8%, and 23.2%, respectively).

Tables 2a and 2b show the 10 top-rank-ing diagnoses exhibiting upward, down-ward, or bimodal trends across the three surveillance periods, by IR differences, among males and females, respectively. For upward trends, mental health disor-ders were the most common type of diag-nosis in the 10 top-ranking diagnoses for both sexes, and musculoskeletal conditions were more common in females than males. However, it should be noted that the sub-stantial increase in the incidence rates of adjustment reaction over time was largely driven by diagnoses of posttraumatic stress disorder (ICD-9: 309.81), which increased from 1.1% to 17.9% of the total 309.* diag-noses among males and from 3.1% to 16.4% among females from the first to last period (data not shown). The downward-trending top 10 list comprised mostly respiratory infections/diseases, and injuries/poison-ings were more common in females than males. Lastly, for bimodal trends, muscu-loskeletal conditions were the most com-mon conditions for both sexes in their top 10 lists.

June 2018 Vol. 25 No. 6 MSMR Page 13

Mental health disorders dramatically increased in total encounter burden and were the most common diagnoses exhibit-ing upward trends in incidence across the careers of those separating from the U.S. Armed Forces. The types of mental health disorders exhibiting upward trends in this study for both sexes—“adjustment reac-tion” (ICD-9: 309), “anxiety, dissociative, and somatoform disorders” (ICD-9: 300), and “depressive disorder, not elsewhere classified” (ICD-9: 311)—also were among the most common incident mental health disorder categories including adjustment disorders, depressive disorders, and anxi-ety disorders among all service members during this time period.38 These upward trends in incidence seem counter to trends observed in the civilian population in which most mental health disorders occur during childhood or adolescence.39 How-ever, most (94.2%) of the surveillance pop-ulation in this study were aged 17–34 years, so this finding might not be surprising,

especially as most of the more common disorders diagnosed during military ser-vice (e.g., mood disorders, other anxiety disorders, and substance use disorders)38 are diagnosed after age 18.39 Additionally, military service has many unique stress-ors,40-51 particularly combat and trauma exposure,40-42,44,47-49 which might explain these trends.

Two MSMR studies published in 2010 both found a pronounced increase in the incidence rates of illness and injury-related diagnoses within 6 months of retirement, compared to 12–18 months before retire-ment.15,16 However, mental health con-ditions were neither among the 18 most frequent illnesses/injuries diagnosed among retirees16 nor in the top 25 diag-noses by incident rate differences when comparing retirees versus “pre-retirees” or retirees versus “retirement eligible.”15 These are striking differences compared to this study’s finding of a growing pro-portionate encounter burden across time

in service and upward trends in mental health–related ICD codes for separating service members. There are several poten-tial reasons for this difference. First, this study did not distinguish between vol-untarily separating service members and those being medically separated. Many mental health conditions, especially those lasting longer than a year, requiring treat-ment, and/or impacting duty, do not meet retention standards,11-13 and mental health disorders have been found to be the leading category of discharge diagnoses in men and the second leading category in women.52 Service members reaching retirement are likely among the healthiest overall ser-vice members across time, and this could be the reason for the observed differences between the two populations.

The findings of this report should be interpreted in light of several important limitations. First, because this study did not focus on a single disease or subset of diseases, it would not have been feasible to

F I G U R E . Proportion of total outpatient and inpatient encounters in three surveillance periods,a by sex, by burden of disease major category,23,24 active component, U.S. Armed Forces, 30 September 2000 through 30 September 2015

aFirst, middle, and last 6 months of service for all active component service members separating with at least 48 months of continuous service at the time of separation and with-out any deployment days in the surveillance periods or breaks in service

Figure. Proportion of total outpatient and inpatient encounters in three surveillance periods,a by burden of disease major category,23,24 active component, U.S. Armed Forces, 30 September 2000 through 30September 2015

aFirst, middle, and last 6 months of service for all active component service members separating with at least 48 months of continuous service at the time of separation and without any deployment days in the surveillance periods or breaks in service

Males

Females

First 6 months of service Middle 6 months of service Last 6 months of service

MSMR Vol. 25 No. 6 June 2018 Page 14

T A B L E 2 a . Top 10 diagnoses exhibiting upward,a downward,b or bimodal trendsc across three surveillance periodsd in separating service members' careers, by incidence rate (IR) differences, males only, U.S. Armed Forces, 30 September 2000 through 30 September 2015

Trend ICD-9 code

Disease burden category Diagnosis

IR per 1,000 persons

FirstNo. of

incident cases

MiddleNo. of

incident cases

LastNo. of

incident cases

Upw

ard

724 Musculoskeletal diseases Other and unspecified disorders of back 32.7 1,244 39.3 1,253 80.2 1,909

799 Signs and symptoms Other ill-defined and unknown causes of mor-bidity and mortality 28.5 1,084 50 1,506 75.3 1,605

309 Mental disorders Adjustment reaction 7.1 271 21.9 772 52.1 1,521

327 Neurologic conditions Organic sleep disorders 0.1 3 4.3 163 44.2 1,538

300 Mental disorders Anxiety, dissociative and somatoform disorders 2.7 101 11.7 429 41.2 1,306

311 Mental disorders Depressive disorder, not elsewhere classified 1.9 72 8.8 326 27.8 930

338 Neurologic conditions Pain, not elsewhere classified 0.8 29 8.2 308 23.4 784

388 Sense organ diseases Other disorders of ear 5.2 197 6 222 27.5 959

796 Signs and symptoms Other nonspecific abnormal findings 6.9 261 13.9 500 25.4 822

307 Mental disorders Special symptoms or syndromes not elsewhere classified 3.4 131 6.1 226 21.8 751

Dow

nwar

d

465 Respiratory infections Acute upper respiratory infections of multiple or unspecified sites 281.7 10,729 28.5 661 26.7 533

460 Respiratory infections Acute nasopharyngitis (common cold) 108.5 4,132 6.4 210 5.9 184

462 Respiratory infections Acute pharyngitis 77.9 2,966 20.5 645 17.6 495

079 Infectious and parasitic diseases

Viral and chlamydial infection in conditions clas-sified elsewhere and of unspecified site 69 2,628 13.3 440 10.7 330

466 Respiratory infections Acute bronchitis and bronchiolitis 51.5 1,962 5.7 201 4.6 156

844 Injury and poisoning Sprains and strains of knee and leg 48.8 1,857 16.3 541 10.4 316

682 Skin diseases Other cellulitis and abscess 44.2 1,683 11 379 9.3 303

845 Injury and poisoning Sprains and strains of ankle and foot 48.9 1,862 20.2 656 15 437

490 Respiratory diseases Bronchitis, not specified as acute or chronic 35.3 1,345 4.8 172 4.8 165

704 Skin diseases Diseases of hair and hair follicles 36.1 1,376 8.6 305 7.6 255

Bim

odal

367 Sense organ diseases Disorders of refraction and accommodation 257.3 9,800 47.1 985 66.2 1,095

719 Musculoskeletal diseases Other and unspecified disorders of joint 125.4 4,778 65.9 1,675 106.2 1,780

780 Signs and symptoms General symptoms 54.4 2,071 30.5 975 78.5 1,924

372 Sense organ diseases Disorders of conjunctiva 72.4 2,758 10.5 351 12.7 399

729 Musculoskeletal diseases Other disorders of soft tissues 73.3 2,791 30.3 943 44.7 1,116

486 Respiratory infections Pneumonia, organism unspecified 59.3 2,260 2.6 91 3.1 107

786 Signs and symptoms Symptoms involving respiratory system and other chest symptoms 50.4 1,919 26.4 856 53.4 1,414

726 Injury and poisoning Peripheral enthesopathies and allied syndromes 40.5 1,541 17.8 602 32.7 961

784 Headache Symptoms involving head and neck 36.4 1,385 16.5 557 31.2 927

728 Musculoskeletal diseases Disorders of muscle ligament and fascia 33.2 1,265 15.6 540 30.5 925

aDiagnoses showing any increase in the IR between both the first & middle and middle & last surveillance periods, ranked by the total difference between the last & first period, from largest to smallestbDiagnoses showing any decrease in the IR between both the first & middle and middle & last surveillance periods, ranked by the total difference between the last & first period, from largest to smallestcDiagnoses showing any decrease in the IR between the first & middle surveillance period and any increase between the middle & last period, ranked by the sum of the abso-lute value of the difference between the first & middle period and the middle & last period, from largest to smallestdFirst, middle, and last 6 months of service for each separating service member

June 2018 Vol. 25 No. 6 MSMR Page 15

T A B L E 2 b. Top 10 diagnoses exhibiting upward,a downward,b or bimodal trendsc across three surveillance periodsd in separating service members' careers, by incidence rate (IR) differences, females only, U.S. Armed Forces, 30 September 2000 through 30 September 2015

Trend ICD-9 code

Disease burden category Diagnosis

IR per 1,000 persons

FirstNo. of

incident cases

MiddleNo. of

incident cases

LastNo. of

incident cases

Upw

ard

309 Mental disorders Adjustment reaction 18.0 131 44.6 263 76.3 324

300 Mental disorders Anxiety, dissociative and somatoform disorders 9.1 66 33.3 213 65.8 320

799 Signs and symptoms Other ill-defined and unknown causes of mor-bidity and mortality 64.4 468 89.9 391 111.5 279

724 Musculoskeletal diseases Other and unspecified disorders of back 62.2 452 80.6 394 105.5 325

311 Mental disorders Depressive disorder, not elsewhere classified 6.6 48 27 175 47.2 249

338 Neurologic conditions Pain, not elsewhere classified 2.1 15 17 119 37.9 223

346 Headache Migraine 9.8 71 22.4 145 42.5 235

723 Musculoskeletal diseases Other disorders of cervical region 8.3 60 19.9 133 39 221

739 Musculoskeletal diseases Nonallopathic lesions not elsewhere classified 4.3 31 19.2 129 32 187

327 Neurologic conditions Organic sleep disorders 0.6 4 5.2 37 24.8 167

Dow

nwar

d

465 Respiratory infections Acute upper respiratory infections of multiple or unspecified sites 292.5 2,127 58.7 203 53.3 136

460 Respiratory infections Acute nasopharyngitis [common cold] 155.1 1,128 19.2 106 13.6 68

079 Infectious and parasitic dis-eases

Viral and chlamydial infection in conditions clas-sified elsewhere and of unspecified site 128.0 931 42 214 25.1 107

844 Injury and poisoning Sprains and strains of knee and leg 90.9 661 14.4 84 11.8 63

462 Respiratory infections Acute pharyngitis 104.8 762 45.7 227 37.6 149

845 Injury and poisoning Sprains and strains of ankle and foot 75.8 551 19 112 13.5 71

625 Genitourinary diseases Pain and other symptoms associated with female genital organs 100.5 731 48.2 246 40.2 158

919 Injury and poisoning Superficial injury of other multiple and unspecified sites 62.6 455 5.9 39 2.8 18

795 Signs and symptoms Other and nonspecific abnormal histological and immunological findings 71.5 520 39.2 211 25.9 122

490 Respiratory diseases Bronchitis, not specified as acute or chronic 43.3 315 11 71 8.4 51

Bim

odal

719 Musculoskeletal diseases Other and unspecified disorders of joint 269.2 1,957 74.7 264 120.5 262

367 Sense organ diseases Disorders of refraction and accommodation 283.9 2,064 84.3 244 104.3 200

269 Nutritional disorders Other nutritional deficiencies 170.7 1,241 2 12 3.2 19

729 Musculoskeletal diseases Other disorders of soft tissues 178.4 1,297 50.4 237 72.4 244

726 Injury and poisoning Peripheral enthesopathies and allied syndromes 82.4 599 28.3 164 40.6 197

786 Signs and symptoms Symptoms involving respiratory system and other chest symptoms 85.3 620 44.4 229 69 265

733 Musculoskeletal diseases Other disorders of bone and cartilage 65.7 478 13.6 86 16.7 97

728 Musculoskeletal diseases Disorders of muscle ligament and fascia 51.4 374 27.2 163 51.9 250

784 Headache Symptoms involving head and neck 75.6 550 49.9 259 61.8 238

477 Respiratory diseases Allergic rhinitis 57.4 417 26.6 154 33.6 166

aDiagnoses showing any increase in the IR between both the first & middle and middle & last surveillance periods, ranked by the total difference between the last & first period, from largest to smallestbDiagnoses showing any decrease in the IR between both the first & middle and middle & last surveillance periods, ranked by the total difference between the last & first period, from largest to smallestcDiagnoses showing any decrease in the IR between the first & middle surveillance period and any increase between the middle & last period, ranked by the sum of the abso-lute value of the difference between the first & middle period and the middle & last period, from largest to smallestdFirst, middle, and last 6 months of service for each separating service member

MSMR Vol. 25 No. 6 June 2018 Page 16

apply or develop case definitions for every possible diagnosis. However, for many conditions, a single ICD-9 code may not represent a true or final diagnosis, and many encounters might have been screen-ing encounters or have been coded incor-rectly. These possibilities represent a source of misclassification bias. Second, “trends” were defined in absolute terms (i.e., any change between periods), rather than by relative or minimum percent changes, and defining “trends” differently likely would have produced a different set of results.

Another limitation is that the study population consisted only of those service members separating in a specific 1-year window. Those separating in 1 year might be different in terms of demographic char-acteristics and health from those separat-ing in another year. For instance, the time frame chosen for this study, October 2014 through September 2015, is 5–7 years fol-lowing one of the worst recessions in U.S. history and the drawdown of forces in Iraq and Afghanistan. The individuals who chose to leave military service during this study’s surveillance window might be con-siderably different from those who left dur-ing other time periods; thus, findings from this study may not be generalizable to pre-vious or future separating service mem-bers. Additionally, the surveillance periods were only 6 months in length. Although the length of this period might not be of par-ticular concern for males because of their larger sample size, this short time frame might not have captured a representative picture of disease in the relatively smaller sample of females. Findings were strati-fied only by sex while many other potential confounders likely exist such as age, race/ethnicity, branch of service, military rank/grade, and military occupation. Lastly, ser-vice members who were deployed dur-ing any of the surveillance windows were excluded to allow for equal surveillance opportunity across the three time periods. However, this exclusion may have intro-duced selection bias given that service members who deploy tend to be “healthier” than those who do not deploy.53

This study appears to be one of the first to examine diagnostic trends over specified time points during the careers of individual service members, and several trends were

identified that could offer opportunities for preventive interventions. Compared to prior studies, separating service mem-bers seem to be different from those retir-ing with respect to the incidence of medical conditions prior to leaving service. Volun-tarily separating service members without disability have more difficulty accessing VA healthcare than retiring individuals or those who are medically separated.37 If fur-ther studies show a significant burden of disease among voluntarily separating ser-vice members who are not accessing VA healthcare, this finding could warrant a review of the transition and compensation process when separating service members move from active duty to civilian life.

Author affiliations: Department of Pre-ventive Medicine and Biostatistics, Uni-formed Services University of the Health Sciences, Bethesda, MD (Capt Uptegraft); Armed Forces Health Surveillance Branch, Defense Health Agency, Silver Spring, MD (Dr. Stahlman).

Acknowledgments: The authors thank COL (Dr.) P. Ann Loveless for her project guidance and mentorship; CDR (Dr.) Shawn Clau-sen, Dr. Margaret Venuto, and Dr. D. Wil-liam White for their thoughtful insights; Dr. Saixia Ying for fulfilling this project's data request; and Ms. Amaris Thurston for her administrative support.

Disclaimer: The contents of this publication are the sole responsibility of the authors and do not necessarily reflect the views, asser-tions, opinions, or policies of the Uniformed Services University of the Health Sciences (USU), the Department of Defense (DoD), or the Departments of the Army, Navy, or Air Force. Mention of trade names, com-mercial products, or organizations does not imply endorsement by the U.S. Government.

R E F E R E N C E S

1. Centers for Disease Control and Prevention. Summary Health Statistics: National Health Inter-view Survey, 2015. Atlanta, GA: National Center for Health Statistics; 2015.2. McDermott KW, Elixhauser A, Sun R. Health-care Cost and Utilization Project Statistical Brief No. 225: Trends in Hospital Inpatient Stays in the

United States, 2005–2014. Rockville, MD: Agency for Healthcare Research and Quality; 2017.3. Moore BJ, Stocks C, Owens PL. Healthcare Cost and Utilization Project Statistical Brief No. 227: Trends in Emergency Department Visits, 2006–2014. Rockville, MD: Agency for Healthcare Research and Quality; 2017.4. Pfuntner A, Wier LM, Stocks C. Healthcare Cost and Utilization Project Statistical Brief No. 162: Most Frequent Conditions in U.S. Hospitals, 2011. Rockville, MD: Agency for Healthcare Re-search and Quality; 2013.5. Regitz-Zagrosek V. Sex and gender differenc-es in health: science and society series on sex and science. EMBO Reports. 2012;13(7):596–603.6. Nonfatal occupational injuries and illnesses requiring days away from work, 2015 [press re-lease]. Washington, DC: U.S. Bureau of Labor Statistics; 2016.7. Weiss AJ, Wier LM, Stocks C, Blanchard J. Healthcare Cost and Utilization Project Statistical Brief No. 174: Overview of Emergency Depart-ment Visits in the United States, 2011. Rockville, MD: Agency for Healthcare Research and Quality; 2014.8. U.S. Department of Defense. DoDI 6130.03: Medical Standards for Appointment, Enlistment, or Induction in the Military Services. Washington, DC: USD(P&R); 2012.9. Seltzer CC, Jablon S. Effects of selection on mortality. Am J Epidemiol. 1974;100(5):367–372.10. Kang HK, Bullman TA. Mortality among U.S. veterans of the Persian Gulf War. N Engl J Med. 1996;335(20):1498–1504.11. U.S. Department of the Air Force. Air Force In-struction 48-123: Medical Examinations and Stan-dards. Washington, DC: AF/SG3; 2013.12. U.S. Department of the Army. Army Regulation 40-501: Standards of Medical Fitness. Washington, DC: HQ/USA; 2017.13. U.S. Department of the Navy. Navy Medicine P-117: Manual of the Medical Department, Ch. 15, Medical Examinations. Washington, DC: Bureau of Medicine and Surgery; 2017.14. U.S. Department of Veterans Affairs. Compen-sation. 2017; https://www.benefits.va.gov/compen-sation/. Accessed on 8 November 2017.15. Armed Forces Health Surveillance Center. Ill-ness and injury diagnoses within six months before retirement after 20 or more years of active service, active component, U.S. Armed Forces, 2000–2009. MSMR. 2010;17(10):2–6.16. Armed Forces Health Surveillance Center. Numbers, proportions, and natures of conditions that are diagnosed for the first time within six months before retirement, active component, U.S. Armed Forces, 2003–2009. MSMR. 2010;17(12):2–5.17. Defense Business Board. Report to the Sec-retary of Defense: Modernizing the Military Retire-ment System. October 2011.18. Mullen MG. CJCS Guide to the Chairman's Readiness System. 2010.19. Woodson J. Congressional testimony of the Assistant Secretary of Defense (Health Affairs) and Surgeons General of the military departments. House Armed Services Committee; 2015.20. Aeromedical Services Information Management System. U.S. Department of the Air Force; 2017.21. Medical Operational Data System. U.S. De-partment of the Army; 2017.22. Medical Readiness Reporting System. U.S. Department of the Navy; 2017.

June 2018 Vol. 25 No. 6 MSMR Page 17

23. Armed Forces Health Surveillance Branch. Ab-solute and relative morbidity burdens attributable to various illnesses and injuries, active component, U.S. Armed Forces, 2016. MSMR. 2017;24(4):2–8.24. Murray CJL, Lopez AD. The Global Burden of Disease: A Comprehensive Assessment of Mortal-ity and Disability from Diseases, Injuries, and Risk Factors in 1990 and Projected to 2020. Published by the Harvard School of Public Health on behalf of the World Health Organization and the World Bank; 1996.25. Breese BB, Stanbury J, Upham H, Calhoun AJ, Van Buren RI, Kennedy AS. Influence of crowding on respiratory illness in a large naval training sta-tion. War Med. 1945;7:143–146.26. Cohen S, Tyrrell DA, Smith AP. Psychological stress and susceptibility to the common cold. N Engl J Med. 1991;325(9):606–612.27. Russell KL. Respiratory Infections in Military Recruits. Washington, DC: TMM Publications; 2006.28. O'Shea MK, Wilson D. Respiratory infections in the military. J R Army Med Corps. 2013;159: 181–189.29. Gleeson M. Immune function in sport and exer-cise. J Appl Physiol. 2007;103:693–699.30. Sanchez JL, Cooper MJ, Myers CA, et al. Respiratory infections in the U.S. military: Re-cent experience and control. Clin Microbiol Rev. 2015;28(3):743–800.31. DeFroda SF, Cameron KL, Posner M, Kriz PK, Owens BD. Bone stress injuries in the military: di-agnosis, management, and prevention. Am J Or-thop (Belle Mead NJ). 2017;46(4):176–183.32. Kaufman KR, Brodine S, Shaffer R. Military training-related injuries: surveillance, research, and prevention. Am J Prev Med. 2000;18(3S):54–63.33. Knapik JJ, Canham-Chervak M, Hauret K, Ho-edebecke E, Laurin MJ, Cuthie J. Discharges dur-ing U.S. Army basic training: injury rates and risk factors. Mil Med. 2001;166(7):641–647.34. Nye NS, Pawlak MT, Webber BJ, Tchandja JN,

Milner MR. Description and rate of musculoskeletal injuries in Air Force basic military trainees, 2012–2014. J Athl Train. 2016;51(11):858–865.35. Rauh MJ, Macera CA, Trone DW, Shaffer RA, Brodine SK. Epidemiology of stress fracture and lower-extremity overuse injury in female recruits. Med Sci Sports Exerc. 2006;38(9):1571–1577.36. Wentz L, Liu P-Y, Haymes E, Ilich JZ. Females have a greater incidence of stress fracture than males in both military and athletic populations: a systematic review. Mil Med. 2011;176(4):420–430.37. U.S. Department of Veterans Affairs. Health Benefits. 2016; https://www.va.gov/healthbenefits/resources/priority_groups.asp. Accessed 15 De-cember 2017.38. Armed Forces Health Surveillance Branch. Mental disorders and mental health problems, ac-tive component, U.S. Armed Forces, 2000–2011. MSMR. 2012;19(6):11–17.39. Kessler RC, Amminger GP, Aguilar-Gaxiola S, Alonso J, Lee S, Ustun TB. Age of onset of mental disorders: a review of recent literature. Curr Opin Psychiatry. 2007;20(4):359–364.40. Hoge CW, Castro CA, Messer SC, McGurk D, Cotting DI, Koffman RL. Combat duty in Iraq and Afghanistan, mental health problems, and barriers to care. N Engl J Med. 2004;351(1):13–22.41. Bryan CJ, Hernandez AM, Allison S, Clem-ans T. Combat exposure and suicide risk in two samples of military personnel. J Clin Psychol. 2013;69(1):64–77.42. Prigerson HG, Maciejewski PK, Rosenheck RA. Population attributable fractions of psychiat-ric disorders and behavioral outcomes associated with combat exposure among U.S. men. Am J Pub-lic Health. 2002;92(1):59–63.43. Ursano RJ, Holloway HC, Jones DR, Rodri-guez AR, Belenky GL. Psychiatric care in the mili-tary community: famly and military stressors. Hosp Community Psychiatry. 1989;40(12):1284–1289.44. Cesur R, Sabia JJ, Tekin E. The psychological

costs of war: military combat and mental health. J Health Econ. 2013;32(1):51–65.45. Hom MA, Stanley IH, Schneider ME, Joiner TE Jr. A systematic review of help-seeking and men-tal health service utilization among military service members. Clin Psychol Rev. 2017;53:59–78.46. Kessler RC, Heeringa SG, Stein MB, et al. Thirty-day prevalence of DSM-IV mental disorders among non-deployed soldiers in the US Army: Re-sults from the Army study of assess resilience in servicemembers (Army STARRS). JAMA Psychia-try. 2014;71(5):504–513.47. Moore TM, Risbrough VB, Baker DG, et al. Ef-fects of military service and deployment on clinical symptomatology: the role of trauma exposure and social support. J Psychiatr Res. 2017;95:121–128.48. Wells TS, Leardmann CA, Fortuna SO, et al. A prospective study of depression following combat deployment in support of the wars in Iraq and Af-ghanistan. Am J Public Health. 2010;100(1):90–99.49. Nguyen S, Leardmann CA, Smith B, et al. Is military deployment a risk factor for mater-nal depression? J Womens Health (Larchmt). 2013;22(1):9–18.50. Negrusa B, Negrusa S. Home front: post-de-ployment mental health and divorces. Demogra-phy. 2014;51(3):895–916.51. Drummet AR, Coleman M, Cable S. Military families under stress: implications for family life education. Fam Relat. 2003;52(3):279–287.52. Hoge CW, Lesikar SE, Guevara R, et al. Men-tal disorders among U.S. military personnel in the 1990s: association with high levels of health care utilization and early military attrition. Am J Psychia-try. 2002;159(9):1576–1583.53. Bollinger MJ, Schmidt S, Pugh JA, Parsons HM, Copeland LA, Pugh MJ. Erosion of the healthy soldier effect in veterans of U.S. military ser-vice in Iraq and Afghanistan. Popul Health Metr. 2015;13(8).

MSMR Vol. 25 No. 6 June 2018 Page 18

During 2013–2017, a total of 1,788 active component service members received incident diagnoses of one of the eating disorders: anorexia ner-vosa (AN), bulimia nervosa (BN) or “other/unspecified eating disorder” (OUED). The crude overall incidence rate of any eating disorder was 2.7 cases per 10,000 person-years. Of the case-defining diagnoses, OUED and BN accounted for 46.4% and 41.8% of the total incident cases, respectively. The overall incidence rate of any eating disorder among women was more than 11 times that among men. Overall rates were highest among service members in the youngest age groups (29 years or younger). Crude annual incidence rates of total eating disorders increased steadily between 2013 and 2016, after which rates decreased slightly. Results of the current study suggest that service members likely experience eating disorders at rates that are com-parable to rates in the general population, and that rates of these disorders are potentially rising among service members. These findings underscore the need for appropriate prevention and treatment efforts in this population.

Diagnoses of Eating Disorders, Active Component Service Members, U.S. Armed Forces, 2013–2017Valerie F. Williams, MA, MS; Shauna Stahlman, PhD, MPH; Stephen B. Taubman, PhD

Eating disorders are characterized by significant and persistent distur-bances of eating that are associated

with increased psychopathology, serious physical health problems, impaired psy-chosocial functioning, and reduced qual-ity of life.1-3 Moreover, eating disorders represent a considerable economic burden in terms of work productivity loss, health-care resource utilization, and healthcare costs.4,5 The International Classification of Diseases, Tenth Revision, Clinical Modifi-cation (ICD-10) includes four broad cat-egories of eating disorder types: anorexia nervosa (AN), bulimia nervosa (BN), “other eating disorders,” and “eating disor-ders, unspecified.” Both the types of con-ditions included in these categories and the diagnostic criteria for the specific dis-orders have changed over time. The diag-nostic criteria for these conditions draw on the American Psychiatric Associa-tion Diagnostic and Statistical Manual of

Mental Disorders (DSM-V) classification and are summarized in Table 1.6 Eating disorders are not associated with loss of appetite, are non-organic in origin (i.e., not caused by a known physical illness), and are not directly attributable to other mental disorders.6

The prevalence of eating disorders is generally elevated among young females7,8

and in high-income countries,9,10 possibly attributable to sociocultural and economic factors. In the U.S., eating disorders affect members of all race/ethnicity groups.2,11 Estimates of the prevalence of these disor-ders in the general population vary widely, depending on study methods and popula-tions.12 In a nationally representative U.S. sample, lifetime prevalence estimates of AN and BN were 0.9% and 1.5% among women, and 0.3% and 0.5%, respectively, among men.2

Published studies of the prevalence of eating disorders among U.S. military

members have used a variety of assess-ment methods and have yielded a range of estimates.13-23 Studies of U.S. military pop-ulations that diagnosed eating disorders using clinical interviews reported preva-lence estimates that are generally com-parable to or higher than those obtained from studies of the U.S. general popula-tion, with approximately 5%–8% of ser-vice women and 0.1% of service men diagnosed with an eating disorder.15,16,22 Studies that employed validated eating disorder screening instruments in military populations have described prevalence estimates for AN, BN, and eating disorder not otherwise specified (EDNOS) of 1.1%, 8.1%–12.5%, and 36.0%–62.8% among women, respectively, and 2.5%, 6.8%, and 40.8% among men.13,14,18 Lower estimates were obtained from a study of U.S. service members based on eating disorder diag-noses recorded during hospitalizations and outpatient healthcare encounters (AN, BN, and EDNOS: 0.25%, 0.79%, and 0.72%, respectively, among women, and 0.01%, 0.02%, and 0.03% among men).21

By current Department of Defense (DoD) policy, a diagnosis of AN, BN, or an unspecified eating disorder lasting lon-ger than 3 months and occurring after age 13 is medically disqualifying for acces-sion into military service.24 Moreover, service members affected by eating dis-orders that are unresponsive to treatment and/or interfere with the satisfactory per-formance of their military duties may be referred to a medical evaluation board and may possibly be separated from service.25

Among military populations, several factors could increase risk of developing an eating disorder. Military members are subject to strict service-specific regula-tions regarding physical fitness and weight requirements and their lifestyles are reg-imented.26 It is well recognized that fac-tors that increase emphasis on weight and

June 2018 Vol. 25 No. 6 MSMR Page 19

shape elevate the risk of eating disorders among both women and men.27 Service members’ exposure to potentially trau-matic experiences and their relatively high rates of mental health disorders also may put them at increased risk of developing eating disorders.28,29 In addition, given the increase in the annual prevalence of diag-noses of clinical overweight among U.S. active component service members during 2011–2015,30 eating-disordered behaviors may develop as service men and women attempt to lose or control their weight. Finally, the changing demographics in the military (women are a rapidly grow-ing segment of U.S. military populations) further highlight the need for continued investigation of eating disorders among service member populations.

In 2014, the MSMR reported the over-all and annual incidence rates of AN, BN, and EDNOS among active component service members during 2004–2013.31 That report documented that, through-out the 10-year period, annual incidence rates declined slightly for each disorder and for all three types combined. The cur-rent report updates this earlier work by describing the incidence of diagnoses of AN, BN, and “other/unspecified eating disorders” among active component ser-vice members during 2013–2017.

M E T H O D S

The surveillance period was 1 Janu-ary 2013 through 31 December 2017. The surveillance population consisted of active component service members of the U.S. Army, Navy, Air Force, or Marine Corps who served at any time during the surveil-lance period. All data used to determine incident eating disorder–specific diagnoses were obtained from electronic records rou-tinely maintained in the Defense Medical Surveillance System (DMSS). These records document both hospitalizations and ambu-latory encounters of active component ser-vice members of the U.S. Armed Forces in fixed (i.e., not deployed or at sea) medi-cal facilities of the Military Health System (MHS) and civilian treatment facilities in the purchased care system.

In the current study, an incident case of one of the three eating disorders of inter-est (AN, BN, or other/unspecified eating disorders [OUEDs]) was defined by the presence of any qualifying ICD-9 or ICD-10 diagnosis code in the 1st or 2nd diag-nostic position of a hospitalization record or in the 1st diagnostic position of a record of an outpatient medical encounter.32 Case-defining diagnoses were AN (ICD-9: 307.1; ICD-10: F50.0*), BN (ICD-9: 307.51; ICD-10: F50.2), and OUED (ICD-9: 307.50, 307.59; ICD-10: F50.8, F50.81, F50.89, F50.9) (Table 1).32 The incidence date was considered the date of the first hospitaliza-tion or outpatient medical encounter that included a case-defining diagnosis of an eating disorder.

For summary purposes, each affected service member could be counted as a case of only one of the three types of eating dis-orders once during the surveillance period. To this end, if a service member received more than one eating disorder–specific diagnosis, AN and BN were prioritized over OUED. If an individual received diagnoses of both AN and BN, the diagnosis recorded first was prioritized over subsequent diag-noses. Individuals were classified as OUED cases only if they were not diagnosed with either AN or BN. Service members with case-defining diagnoses before the start of the surveillance period were excluded from the incidence analysis because they were not considered at risk of incident (i.e., first-ever) diagnoses of eating disorders.

Prevalence of the diagnoses of each of the three types of eating disorder was esti-mated for each year in the 5-year surveil-lance period. The numerator for prevalence calculations consisted of those individuals identified as incident cases of an eating dis-order in a given year or in a previous year and who also had a healthcare encounter for any eating disorder type during that year. The denominator for prevalence cal-culations consisted of the total number of active component service members who served at least 1 day of the given year. Prev-alence estimates were calculated for each of the three eating disorders of interest (AN, BN, and OUED) as the number of preva-lent cases per 10,000 active component ser-vice members.

R E S U L T S

During the 5-year surveillance period, a total of 1,788 active component service members received incident diagnoses of eating disorders, for a crude overall inci-dence rate of 2.7 cases per 10,000 person-years (p-yrs) (Table 2). Of the case-defining diagnoses, OUED and BN accounted for 46.4% and 41.8% of the total incident cases, respectively. Less than one-eighth (11.9%) of the total incident cases of eating disor-der were attributable to AN. In regard to all eating disorders, more than two-thirds (67.5%) of incident cases affected females, and the overall incidence rate among women (11.9 cases per 10,000 p-yrs) was more than 11 times that among men (1.0 per 10,000 p-yrs) (Table 2). Crude over-all incidence rates of AN, BN, and OUED among women were 15.7, 15.5, and 8.3 times the rates among men, respectively.

The distributions of incident diagno-ses of the three types of eating disorders by demographic characteristics are pre-sented in Table 3. For both sexes, overall incidence rates were highest among service members in the youngest age groups (29 years or younger) and rates decreased with increasing age. Compared to their respec-tive female counterparts, overall rates were highest among non-Hispanic white ser-vice women (15.8 cases per 10,000 p-yrs), Marine Corps members (20.4 cases per 10,000 p-yrs), junior enlisted or junior offi-cers (16.0 and 11.4 cases per 10,000 p-yrs, respectively), and those in combat-specific occupations (17.2 cases per 10,000 p-yrs). Of note, the overall incidence rate of all eat-ing disorders among female Marine Corps members was nearly twice that among female Army members. Among men, over-all rates were highest among Hispanic ser-vice members, those of other/unknown race/ethnicity and non-Hispanic white ser-vice members (1.3, 1.3, and 1.1 cases per 10,000 p-yrs, respectively), compared to those in other race/ethnicity groups (Table 3). Relative to their respective male counter-parts, rates of diagnoses of all eating disor-ders for men were highest among Army or Marine Corps members (1.2 and 1.1 cases per 10,000 p-yrs, respectively), enlisted ser-vice members (1.2 and 1.1 cases per 10,000

MSMR Vol. 25 No. 6 June 2018 Page 20

T A B L E 1 . Summary of diagnostic criteria for anorexia nervosa, bulimia nervosa, and other/unspecified eating disordersa

Eating disorder ICD-9 code ICD-10 code

Anorexia nervosa (AN) 307.1 F50.0*

Persistent restriction of energy intake leading to significantly low body weight (in context of what is minimally expected for age, sex, developmental trajectory, and physical health)

Intense fear of gaining weight/becoming fat or persistent behavior that interferes with weight gain (even though significantly low weight)

Disturbance in the way one's body weight or shape is experienced, undue influence of body shape and weight on self-evaluation, or persistent lack of recognition of the seriousness of the current low body weight AN, unspecified F50.00 Restricting type: Restriction of food intake; use of fasting, diet pills, exercise; no binge eating or purging F50.01 Binge-eating/purging type: Binge eating and purging behavior to lose weight F50.02

Bulimia nervosa (BN) 307.51 F50.2

Recurrent episodes of binge eating. An episode of binge eating is characterized by both:

• Eating, in a discrete period of time (e.g., within any 2-hour period), an amount of food that is definitely larger than most people would eat during a similar time period and under similar circumstances.

• A sense of lack of control over eating during the episode (e.g., a feeling that one cannot stop eating or control what or how much one is eating)

Recurrent inappropriate compensatory behavior to prevent weight gain (e.g.,self-induced vomiting, misuse of laxatives, diuretics, or other medications, fasting, or excessive exercise) is present.

The binge eating and inappropriate compensatory behaviors both occur, on average, at least once a week for 3 months.

Self-evaluation is unduly influenced by body shape and weight.

The disturbance does not occur exclusively during episodes of anorexia nervosa.

Other/unspecified eating disorders (OUED) 307.50, 307.59 F50.8, F50.81, F50.89, F50.9

Other eating disordersb F50.81

Binge eating disorderc is defined as recurring episodes of eating significantly more food in a short period of time (e.g., 2-hour period) than most people would eat under similar circumstances, with episodes marked by feelings of lack of control over eating during the epidsode. The binge-eating episodes are as-sociated with three or more of the following:

• Eating much more rapidly than normal• Eating until feeling uncomfortably full• Eating large amounts of food when not feeling physically hungry• Eating alone because of feeling embarrassed by how much one is eating• Feeling disgusted with oneself, depressed, or very guilty afterwards

Marked distress regarding binge eating is present. The binge eating occurs, on average, at least once a week for 3 months. The binge eating is not associated with recurrent use of inappropriate compensatory behaviors and does not occur exclusively during the course of AN, BN, or Avoidant/Restrictive Food Intake Disorder.

Other specified eating disorderc includes pica in adults, and psychogenic loss of appetite. F50.89

Unspecified eating disorder: This category applies to behaviors that cause clinically significant distress/ impairment of functioning, but do not meet the full criteria of any of the eating disorder criteria. This category may be used by clinicians when they choose not to specify why criteria are not met, including presentations where there may be insufficient information to make a more specific diagnosis (e.g., in emergency room settings). Includes atypical AN and atypical BN.

F50.9

aAmerican Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), Arlington, VA: American Psychiatric Association; 2013.bICD-10 code F50.8 (other eating disorders) was added to the ICD-10 coding system on 1 October 2015 and was changed to a parent code on 1 October 2016.cF50.81 (binge eating disorders) and F50.89 (other specified eating disorder) were added to the ICD-10 coding system on 1 October 2016.

June 2018 Vol. 25 No. 6 MSMR Page 21

p-yrs for E1–E4 and E5–E9, respectively), and those in healthcare occupations (2.3 cases per 10,000 p-yrs).

Crude annual incidence rates of total eating disorders increased steadily from 2.3 cases per 10,000 p-yrs in 2013 to a peak of 3.3 cases per 10,000 p-yrs in 2016 (44.7% increase), after which rates decreased to 2.9 cases per 10,000 p-yrs in 2017 (Figure 1). Annual rates of OUED remained stable during 2013–2014 at 0.9 cases per 10,000 p-yrs and then increased to a peak of 1.8 cases per 10,000 p-yrs in 2016. Crude rates of diagnoses of BN decreased slightly dur-ing the surveillance period, from 1.2 cases per 10,000 p-years in 2013 to 0.9 cases per 10,000 p-yrs in 2017. Annual rates of AN increased slightly from 0.2 cases per 10,000 p-years in 2013 to 0.4 cases per 10,000 p-yrs in 2017. Of note, the crude annual incidence rate of OUED converged with that of BN in 2015, after which annual rates of OUED exceeded those of BN (Figure 1). The increasing trend in crude annual rates of total eating disorders during 2013–2016 was driven largely by increases in OUED among service members of both sexes dur-ing this period (Figure 2, Figure 3). Among male service members, the annual rates of incident diagnoses of BN and AN were relatively stable during the 5-year surveil-lance period. Among service women, crude annual rates of BN declined during the period, with the greatest decrease occur-ring between 2015 and 2017. Rates of AN among service women increased slightly during the period (Figure 3).

The general pattern of period preva-lences of OUED, BN, and AN among active component service members by year during the surveillance period was broadly similar to that observed for the annual incidence rates of these eating disorder types (Figure 4). The peak prevalences for women were as follows: AN, 3.7 cases per 10,000 active component service women in 2016; BN, 10.3 cases per 10,000 in 2014; and OUED, 13.2 cases per 10,000 in 2016. For men, the peak prevalences were as follows: AN, 0.22 cases per 10,000 active component service men in 2016 and 2017; BN, 0.64 cases per 10,000 in 2016; and OUED, 1.5 cases per 10,000 in 2016 (data not shown).

T A B L E 2 . Incident cases and incidence rates, eating disorders, active component, U.S. Armed Forces, 2013–2017

T A B L E 3 . Incident cases and incidence rates, all eating disorders, active component, U.S. Armed Forces, 2013–2017

All eating disorders, total

Anorexia nervosa

Bulimia nervosa

Other/unspecified eating disorders

No. Ratea No. Ratea No. Ratea No. Ratea

Female 1,207 11.9 157 1.6 552 5.5 498 4.9

Male 581 1.0 55 0.1 195 0.4 331 0.6

Total 1,788 2.7 212 0.3 747 1.1 829 1.3

Female:Male RR 11.4 15.7 15.5 8.3

Female % (all cases) 67.5% 74.1% 73.9% 60.1%

RR, rate ratioaRate per 10,000 person-years

Males Females Female:MaleNo. Ratea No. Ratea RR

Total 581 1.0 1,207 11.9 11.4Age group (years)<21 67 1.0 201 14.6 15.121–24 173 1.2 429 16.1 13.125–29 148 1.1 315 12.6 11.130–34 85 0.9 141 8.7 9.235–39 59 0.9 72 6.9 7.440+ 49 0.8 49 5.5 6.7

Race/ethnicityNon-Hispanic white 357 1.1 707 15.8 15.0Non-Hispanic black 67 0.8 182 6.9 8.2Hispanic 99 1.3 162 10.0 7.7Asian/Pacific Islander 14 0.7 44 10.2 15.2American Indian/Alaska Native 2 0.4 15 13.1 36.2Other 42 1.3 97 11.5 8.8

ServiceArmy 263 1.2 415 11.9 9.6Navy 107 0.8 329 11.4 13.8Air Force 115 0.9 315 10.4 11.5Marine Corps 96 1.1 148 20.4 18.4

RankJunior enlisted (E1–E4) 290 1.2 733 16.0 13.2Senior enlisted (E5–E9) 236 1.1 292 8.2 7.6Officers (O1–O3 [W1–W3]) 37 0.6 152 11.4 18.0Officers (O4–O10 [W4–W5]) 18 0.5 30 4.8 10.2

OccupationCombat-specificb 72 0.8 34 17.2 21.8Motor transport 13 0.8 30 9.6 11.7Pilot/air crew 5 0.2 7 4.8 22.6Repair/engineer 169 1.0 229 11.3 11.5Communications/intelligence 134 1.2 377 11.4 9.3Health care 91 2.3 269 13.7 5.9Other/unknown 97 0.9 261 12.2 13.0

RR, rate ratioaRate per 10,000 person-yearsbInfantry/artillery/combat engineering/armor

MSMR Vol. 25 No. 6 June 2018 Page 22

E D I T O R I A L C O M M E N T

Results of the current analysis indi-cate that crude annual incidence rates of total eating disorders increased by 44.7%

between 2013 and 2016 followed by a slight decrease in 2017. Sex-stratified rates showed that the increasing trend in annual rates of total eating disorders during this period was driven largely by increases in OUED among service members of both

sexes. Previous MSMR results showed that crude annual rates of BN were consistently higher than rates of EDNOS and AN.31 However, in the current study, the crude annual rate of OUED (category most similar to EDNOS category) converged with that of BN in 2015. From that point through 2017, crude annual rates of OUED exceeded those of BN. The increase in OUED rates observed in the current analysis is likely due, at least in part, to adjustments to the classification of eating disorders made in the ICD-10 coding system.

In the previous MSMR report, all diag-noses used the ICD-9 classification system in which binge eating disorder (BED) was not specifically described but was included in the EDNOS category.31 On 1 October 2015, code F50.8 (other eating disorders) was added to the ICD-10 coding system. Subsequently, on 1 October 2016, F50.8 was changed to a parent code and codes F50.81 (BED) and F50.89 (other specified eating disorder) were added to the coding system. Another important adjustment was to clarify the body weight criterion for AN. Previously, “minimal normal body weight” was defined as a body weight less than 85% of that expected. Currently, the criterion is “significantly low body weight” in the con-text of what is minimally expected for age, sex, developmental trajectory, and physical health (Table 1).6 In addition, the amenor-rhea criterion under AN was removed.6 For BN, the minimum frequency of binge eat-ing episodes and inappropriate compen-satory behavior was reduced from twice a week to once a week.6

Results of several U.S. studies indicate that BED is one of the most common eat-ing disorders among both sexes.2,33,34 In a U.S. nationally representative sample, life-time prevalence estimates for BED were 3.5% among women and 2.0% among men.2 The increase in crude annual inci-dence rates of OUED among both sexes and their peak in 2016 broadly coincides with the shift to the ICD-10 classification system. Results of some studies suggest that there have been increases in the prev-alence of total eating disorders and of AN over time.35,36 The reasons for this reported increase in prevalences is unclear, but the change in the diagnostic coding of BED and the relaxation of the criteria for AN

F I G U R E 1 . Annual incidence rates of eating disorders, by type, active component, U.S. Armed Forces, 2013–2017

Note: ICD-10 code F50.8 (Other eating disorders) was added to the ICD-10 coding system on 1 October 2015 and was changed to a parent code on 1 October 2016. F50.81 (Binge eating disorders) and F50.89 (Other specified eating disorder) were added to the ICD-10 coding system on 1 October 2016. This likely affected the rates of “other eating disorders” around this time period.

2.3

2.9

1.20.90.9

1.7

0.20.4

0.0

1.0

2.0

3.0

4.0

5.0

2013 2014 2015 2016 2017

p 000,01 rep etar ecnedicnI-y

rs

Total eating disordersBulimia nervosaOther/unspecified eating disordersAnorexia nervosa

0.30

0.83

0.330.27

0.040.12

0.0

0.5

1.0

1.5

2.0

2013 2014 2015 2016 2017

p 000,01 rep etar ecnedicnI-y

rs

Other/unspecified eating disordersBulimia nervosaAnorexia nervosa

Note: ICD-10 code F50.8 (Other eating disorders) was added to the ICD-10 coding system on 1 October 2015 and was changed to a parent code on 1 October 2016. F50.81 (Binge eating disorders) and F50.89 (Other specified eating disorder) were added to the ICD-10 coding system on 1 October 2016. This likely affected the rates of “other eating disorders” around this time period.

F I G U R E 2 . Annual incidence rates of eating disorders, by type, active component males, U.S. Armed Forces, 2013–2017

June 2018 Vol. 25 No. 6 MSMR Page 23

are generally cited as the major reasons for such increases in prevalence over time.12,37

The overall incidence rate of AN diag-noses among women documented here (1.6 cases per 10,000 p-yrs) is comparable to the range of rates cited in Hsu’s review (0.14 to 5.0 cases per 10,000 young women per year); however, prevalence estimates among female service members for AN and BN were much lower.38 The estimates of prevalence yielded by the current analy-sis are otherwise consistent with the pub-lished literature with respect to age group and sex differences and the relative fre-quencies of the three diagnostic catego-ries examined. Both the incidence rates and prevalence estimates in the current analysis are lower than many published estimates from military and civilian pop-ulations.39 However, most of these studies used data from non-military populations and/or employed estimation methods dif-ferent from those used here.

Given that DoD standards preclude entrance into military service for individ-uals with diagnosed eating disorders, it is plausible that the incidence and prevalence of these conditions among service mem-bers are lower than in the civilian popula-tion because eating disorders commonly have their onsets during adolescence (before the age of eligibility for military accession). Nevertheless, the current anal-ysis documents that there are hundreds of new cases of eating disorders diagnosed each year among active component service members; there are likely many other ser-vice members whose conditions went med-ically undetected. The published literature documents that, at least in certain select populations, abnormal eating behaviors occur with surprising frequency among military personnel.13-23 Barlett and Mitch-ell’s systematic review summarizes the lit-erature regarding eating disorders among military and veteran men and women.39

Subgroup-specific results of the cur-rent analysis are consistent with findings in the published literature on eating dis-orders. As expected, based on studies in the U.S. civilian and military populations, incidence rates among women were con-siderably higher than those among men. Female service members accounted for 68% of all diagnosed eating disorders,

even though women account for only 15.9% of active component service mem-bers.40 Similar to results found in rep-resentative samples of military service members,21,23 overall incidence rates of total eating disorders were highest among

non-Hispanic white service members, those in the youngest age groups (29 years or younger), and Marine Corps members.

Results of the current study must be interpreted in the context of several analysis limitations. First, the reliance on

F I G U R E 3 . Annual incidence rates of eating disorders, by type, active component females, U.S. Armed Forces, 2013–2017

Note: ICD-10 code F50.8 (Other eating disorders) was added to the ICD-10 coding system on 1 October 2015 and was changed to a parent code on 1 October 2016. F50.81 (Binge eating disorders) and F50.89 (Other specified eating disorder) were added to the ICD-10 coding system on 1 October 2016. This likely affected the rates of “other eating disorders” around this time period.

F I G U R E 4 . Annual prevalence of eating disorders, by type, active component, U.S. Armed Forces, 2013–2017

Note: ICD-10 code F50.8 (Other eating disorders) was added to the ICD-10 coding system on 1 October 2015 and was changed to a parent code on 1 October 2016. F50.81 (Binge eating disorders) and F50.89 (Other specified eating disorder) were added to the ICD-10 coding system on 1 October 2016. This likely affected the rates of “other eating disorders” around this time period.

6.2

4.04.0

6.0

1.11.8

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

2013 2014 2015 2016 2017

p 000,01 rep etar ecnedicnI-y

rs

Bulimia nervosaOther/unspecified eating disordersAnorexia nervosa

3.3

4.0

2.2

3.1

2.0

1.8

0.470.77

0.0

1.0

2.0

3.0

4.0

5.0

6.0

2013 2014 2015 2016 2017

srebme

m ecivres 000,01 rep sesac fo .oN

Total eating disordersOther/unspecified eating disordersBulimia nervosaAnorexia nervosa

MSMR Vol. 25 No. 6 June 2018 Page 24

diagnoses from records of service mem-bers’ medical encounters undoubtedly resulted in underestimates of the true inci-dence and prevalence of eating disorders among the surveillance population. Per-sons with eating disorders generally avoid seeking medical care, at least initially, either because they do not believe they have a medical problem or because they are embarrassed about their behaviors.33 Individuals with BN or OUED (including BED) are better able to conceal their eat-ing disorders because their body weights and appearances are not suggestive of dis-ordered eating, and their binge eating and compensatory behaviors usually take place in private.33 Service members with these disorders may not have the diagnoses doc-umented in their medical records unless they seek assistance for or experience a serious complication of their conditions.

Another limitation of the current analysis is related to the implementa-tion of MHS GENESIS, the new elec-tronic health record for the MHS. During 2017, medical data from sites that were using MHS GENESIS are not available in DMSS. These sites include Naval Hospital Oak Harbor, Naval Hospital Bremerton, Air Force Medical Services Fairchild, and Madigan Army Medical Center. Therefore, medical encounter and person-time data for individuals seeking care at one of these facilities during 2017 were excluded from the analysis.

Among military personnel, there is reason to believe that some concealment of eating disorders is motivated by con-cerns that discovery and formal diagnosis may influence deployment, promotion,41 or even retention.19 Because of the emaciation that follows extreme weight loss in AN, ser-vice members with this eating disorder are more likely to be noticed by their families, friends, and/or military colleagues and per-suaded to seek medical attention. Among active component service members, deteri-oration of not only physical appearance but also duty performance may serve as triggers for supervisors to refer persons with AN for medical evaluation.19 However, because such medical scrutiny likely follows many months or a few years of weight loss, diag-noses of AN often are documented long after the onset of the disorder.

When AN persists, the debilitating effects have adverse impacts on the physi-cal and mental health and social and occu-pational activities of those affected. In addition, AN that persists or recurs is life threatening. Manos et al. cite studies that estimate crude 10-year mortality rates of 3.3%–5.6% and 20-year rates of 15%–20%.19 Recognition and treatment of AN is essential. In the U.S. Armed Forces, where periodic measurement of service members’ height and weight is common, the detec-tion of a body mass index (BMI) of less than 17.5 kg/m2 should indicate the need for further evaluation.

Because service members affected by BN or OUED usually have BMIs that are in or near the normal range, their appearances may not be indicative of their abnormal eating behaviors. Potential com-plications of BN and OUED that may lead those affected to seek medical care include the consequences of overeating and vomit-ing as well as overuse of laxatives, diuret-ics, and enemas. An extended period of repeated, induced vomiting may result in erosion of dental enamel by the exposure of the teeth to stomach acid.19 Although mortality is a much less frequent outcome of BN and OUED than AN, purging and metabolic abnormalities may be associ-ated with potentially fatal events such as esophageal tears, gastric ruptures, and car-diac arrhythmias.42 Because individuals with eating disorders are at elevated risk of psychiatric comorbidity,1 it is important for military healthcare providers to be vig-ilant for eating disorder symptoms among service members affected by mental health disorders, especially among those in the accession phase of military service.43

Results of the current study suggest that service members likely experience eating disorders at rates that are compara-ble to rates in the general population, and that rates of these disorders are potentially rising among military members. These findings underscore the need for appro-priate prevention and treatment efforts in this population. At stake are the health, well-being, and military operational effec-tiveness of affected service members and their units.

R E F E R E N C E S

1. Pearlstein T. Eating disorders and comorbidity. Arch Womens Ment Health. 2002;4(3):67–78.2. Hudson JI, Hiripi E, Pope HG Jr, Kessler RC. The prevalence and correlates of eating disorders in the National Comorbidity Survey Replication. Biol Psychiatry. 2007;61(3):348–358.3. Jenkins PE, Hoste RR, Meyer C, Blissett JM (2010) Eating disorders and quality of life: a review of the literature. Clin Psychol Rev. 31(1):113–121. 4. Simon J, Schmidt U, Pilling S. The health ser-vice use and cost of eating disorders. Psychol Med. 2005;35(11):1543–1551.5. Ling YL, Rascati KL, Pawaskar M. Direct and indirect costs among patients with binge-eating disorder in the United States. Int J Eat Disord. 2017;50(5):523–532.6. American Psychiatric Association. Feeding and Eating Disorders. In: Diagnostic and Statis-tical Manual of Mental Disorders, Fifth Edition (DSM-5). Arlington, VA: American Psychiatric As-sociation; 2013.7. Favaro A, Ferrara S, Santonastaso P. The spectrum of eating disorders in young women: a prevalence study in a general population sample. Psychosom Med. 2003;65(4):701–708.8. Smink FR, van Hoeken D, Hoek HW. Epide-miology of eating disorders: Incidence, preva-lence and mortality rates. Curr Psychiatry Rep. 2012;14(4):406–414.9. Makino M, Tsuboi K, Dennerstein L. Preva-lence of eating disorders: a comparison of West-ern and non-Western countries. Med Gen Med. 2004;6(3):49.10. Kessler RC, Berglund PA, Chiu WT, et al. The prevalence and correlates of binge eating disorder in the World Health Organization World Mental Health Surveys. Biol Psychiatry. 2013;73(9):904–914.11. Becker AE, Grinspoon SK, Klibanski A, Herzog DB. Eating disorders. N Engl J Med. 1999;340(14):1092–1098.12. Qian J, Hu Q, Wan Y, et al. Prevalence of eating disorders in the general population: a systematic review. Shanghai Arch Psychiatry. 2013;25(4):212–223.13. McNulty PA. Prevalence and contributing fac-tors of eating disorder behaviors in a population of female Navy nurses. Mil Med. 1997;162(10):703–706.14. McNulty PA. Prevalence and contributing fac-tors of eating disorder behaviors in active duty Navy men. Mil Med. 1997;162(11):753–758.15. Lauder TD, Williams MV, Campbell CS, Da-vis G, Sherman R, Pulos E. The female athlete triad: prevalence in military women. Mil Med. 1999;164(9):630–635.16. Lauder TD, Williams MV, Campbell CS, Da-vis GD, Sherman RA. Abnormal eating behav-iors in military women. Med Sci Sports Exerc. 1999;31(9):1265–1271.17. Lauder TD and Campbell CS. Abnormal eating behaviors in female reserve officer training corps cadets. Mil Med. 2001;166(3):264–268.18. McNulty PA. Prevalence and contributing fac-tors of eating disorder behaviors in active duty service women in the Army, Navy, Air Force, and Marines. Mil Med. 2001;166(1):53–58.19. Manos GH, Carlton J, DelaCruz G, Kelley, WA.

June 2018 Vol. 25 No. 6 MSMR Page 25

Eating disorders. In: DeKoning B, ed. Recruit Medi-cine (Textbooks of Military Medicine). Washington, DC: Borden Institute; 2006: pages 357–379.20. Warner C, Warner C, Matuszak T, et al. Dis-ordered eating in entry-level military personnel. Mil Med. 2007;172(2):147–151.21. Antczak AJ, Brininger TL. Diagnosed eating disorders in the U.S. military: a nine year review. Eat Disord. 2008;16(5):363–377.22. Beekley MD, Byrne R, Yavorek T, Kidd K, Wolff J, Johnson M. Incidence, prevalence, and risk of eating disorder behaviors in military academy ca-dets. Mil Med. 2009;174(6):637–641. 23. Jacobson IG, Smith TC, Smith B, et al. Disor-dered eating and weight changes after deployment: Longitudinal assessment of a large US military co-hort. Amer J Epidemiol. 2009;169(4):415–427.24. Department of Defense. Instruction 6130.03, Medical Standards for Appointment, Enlistment, or Induction into the Military Services. Change 1. 30 March 2018. http://www.esd.whs.mil/Portals/54/Documents/DD/issuances/dodi/613003p.pdf. Ac-cessed on 1 May 2018.25. Headquarters, Department of the Army. Army Regulation 40-501. Medical Services. Standards of Medical Fitness. Revision 14 June 2017. https://ar-mypubs.army.mil/epubs/DR_pubs/DR_a/pdf/web/ARN3801_AR40-501_Web_FINAL.pdf. Accessed on 15 May 2018.26. Carlton JR, Manos GH, Van Slyke JA. Anxiety and abnormal eating behaviors associated with cyclical readiness testing in a naval hospital active duty population. Mil Med. 2005;170(8):663–667.

27. Keel PK, Forney KJ. Psychosocial risk factors for eating disorders. Int J Eat Disord. 2013;46(5):433–439.28. Forman-Hoffman VL, Mengeling M, Booth BM, Torner J, Sadler AG. Eating disorders, post-trau-matic stress, and sexual trauma in women veter-ans. Mil Med. 2012;177(10):1161–1168.29. Mitchell KS, Mazzeo SE, Schlesinger MR, Brewerton TD, Smith BN. Comorbidity of partial and subthreshold PTSD among men and women with eating disorders in the national comorbid-ity survey-replication study. Int J Eat Disord. 2012;45(3):307–315.30. Clark LL, Taubman SB. Update: Diagnoses of overweight and obesity, active component, U.S. Armed Forces, 2011–2015. MSMR. 2016;23(9):9–13.31. Armed Forces Health Surveillance Center. Di-agnoses of eating disorders among active compo-nent service members, U.S. Armed Forces, 2004–2013. MSMR. 2014;21(9):8–12.32. Armed Forces Health Surveillance Branch. Surveillance Case Definition. Eating disorders. May 2018. https://www.health.mil/Reference-Cen-ter/Publications/2017/06/01/Eating-Disorders.33. Stice E, Marti CN, Rohde P. Prevalence, inci-dence, impairment, and course of the proposed DSM-5 eating disorder diagnoses in an 8-year prospective community study of young women. J Abnorm Psychol. 2013;122(2):445–457.34. Hudson JI, Coit CE, Lalonde JK, Pope HG Jr. By how much will the proposed new DSM-5 criteria increase the prevalence of binge eating disorder?

Int J Eat Disord. 2012;45(1):139–41.35. Hay PJ, Mond J, Buttner P, Darby A. Eating dis-order behaviors are increasing: findings from two sequential community surveys in South Australia. PloS One. 2008;3(2):e1541.36. Zachrisson HD, Vedul-Kjelsås E, Götestam KG, Mykletun A. Time trends in obesity and eating disorders. Int J Eating Disord. 2008;41(8):673–680.37. Lindvall Dahlgren C, Wisting L, Rø Ø. Feeding and eating disorders in the DSM-5 era: a system-atic review of prevalence rates in non-clinical male and female samples. J Eat Disord. 2017;5:56.38. Hsu LKG. Epidemiology of the eating disorders. Psychiatr Clin North Am. 1996;19(4):681–700.39. Bartlett BA, Mitchell KS. Eating disorders in military and veteran men and women: A systematic review. Int J Eat Disord. 2015;48(8):1057–1069.40. Department of Defense, Office of the Deputy Assistant Secretary of Defense for Military Commu-nity and Family Policy (ODASD (MC&FP)). (2017). 2016 Demographics: Profile of the Military Commu-nity. Washington, DC.41. Greene-Shortridge TM, Britt TW, Castro CA. The stigma of mental health problems in the mili-tary. Mil Med. 2007;172(2):157–161.42. Herzog DB, Copeland PM. Eating disorders. N Engl J Med. 1985;313(5):295–303.43. Johnson WB, Davis K, Gonzalez V. Military accession and disordered eating among women: clinical and policy recommendations. Mil Med. 2014;179(9):936–941.

MSMR Vol. 25 No. 6 June 2018 Page 26

The 2017–2018 influenza season has been a topic of interest in the media and among the general public due

to concerns about the protective nature of the 2017–2018 influenza vaccine. During the Southern Hemisphere’s winter influ-enza season in mid-2017, Australia’s over-all influenza vaccine effectiveness (VE) was surprisingly low at 33% (95% CI, 17%–46%).1 More specifically, Australia reported an influenza A(H3) VE of 10% (95% CI, -16%–31%), which was not statistically sig-nificantly different from zero.1 These find-ings prompted concerns about the prospect of a similarly low VE during the subsequent influenza season in the Northern Hemi-sphere, as Australia and the U.S. selected identical vaccine strains. The Department of Defense (DoD) conducts VE analy-ses to determine the extent of matching between the recommended seasonal vac-cine and the circulating strain. This article reports the results of DoD VE mid-season estimates determined by the Armed Forces Health Surveillance Branch (AFHSB) Air Force (AFHSB-AF) satellite at the U.S. Air Force School of Aerospace Medicine; Naval Health Research Center (NHRC); and the AFHSB.

M E T H O D S

The AFHSB-AF satellite branch is a sentinel site-based program that requests weekly submissions of six to 10 specimens accompanied by a completed question-naire from each site. Vaccination status was verified through immunization records obtained from the Air Force Complete Immunization Tracking Application, medi-cal records from the Aeromedical Services

Information Management System, or self-reported data from the questionnaire. Indi-viduals were considered to be vaccinated if they were vaccinated at least 14 days prior to symptom onset. Those who were vacci-nated less than 14 days prior to symptom onset were excluded from the study.

NHRC’s population included civil-ians who sought care at outpatient clinics near the U.S.–Mexico border through the febrile respiratory illness program.Vaccina-tion status was obtained through medical record reviews and self-report, if available.2 NHRC classified cases and controls to have been vaccinated if symptom onset started 14–180 days after receiving the vaccine.2

AFHSB’s VE study used data obtained via the Defense Medical Surveillance Sys-tem and Navy and Marine Corps Public Health Center. The high vaccination rate is attributable to the fact that annual influ-enza vaccination is required for service members.2

All three VE estimates were derived using a test-negative case-control study design although each organization utilized different study populations (i.e., AFHSB-AF satellite, DoD dependent data; NHRC, civilians near the U.S.–Mexico border; AFHSB, active component service mem-ber data). All studies calculated crude and adjusted VE using odds ratios (ORs) and 95% CIs obtained from multivariable logistic regression models (Table). Statis-tical data analyses were performed using SAS version 9.4 (2013, SAS Institute, Cary, NC). VE was calculated as (1−OR) × 100. AFHSB-AF’s adjustment variables were age group, time period, and geographic region. NHRC’s only adjustment variable was age group. AFHSB’s adjustment variables were age group, sex, month of illness, and 5-year

Brief Report Department of Defense Midseason Vaccine Effectiveness Estimates for the 2017–2018 Influenza SeasonLisa Shoubaki, MPH; Angelia A. Eick-Cost, PhD, ScM; Anthony W. Hawksworth, BS; Zheng Hu, MS; LeeAnne Lynch, MPH; Christopher A. Myers, PhD; Susan Federinko, MD, MPH (Lt Col, USAF)

vaccination status. For summary purposes, vaccine effects were considered statistically significant if 95% CIs around point esti-mates of VE did not include zero.

Inactivated influenza vaccine was the only vaccine type analyzed, because the live, attenuated influenza vaccine was not rec-ommended or used during the 2017–2018 season. Cases were laboratory-confirmed influenza positives and controls were influ-enza test negatives. Influenza positives from the AFHSB-AF satellite and NHRC were confirmed through reverse transcrip-tion polymerase chain reaction (RT-PCR) and/or viral culture, while AFHSB used RT-PCR and/or viral culture as well as pos-itive rapid tests, excluding individuals with rapid test negatives.

R E S U L T S

From 1 October 2017 through 10 Feb-ruary 2018, the AFHSB-AF’s VE study included 1,160 cases and 1,383 controls, with 36% and 47% having been vaccinated, respectively. Overall, the adjusted VE was 51% (95% CI, 41%–59%). The adjusted VE for influenza A(H3N2) was low at 37% (95% CI, 22%–49%) (Figure). Influenza A(H1N1)pdm09 and influenza B had higher adjusted VE estimates of 79% (95% CI, 67%–86%) and 60% (95% CI, 49%–70%), respectively (Figure). Adjusted VE estimates were simi-lar among children (aged 2–17 years) and adults (data not shown).

From 13 November 2017 through 8 January 2018, the NHRC’s VE study included 201 cases and 114 controls, with 13% and 24% having been vaccinated, respectively. For the NHRC’s study, the overall adjusted VE was 55% (95% CI,

June 2018 Vol. 25 No. 6 MSMR Page 27

17%–75%). For influenza A(H3N2), VE was 52% (95% CI, 9%–75%). For influenza B, VE was 63% but not statistically signifi-cant (95% CI, -5%–87%) (Figure).

From 1 December 2017 through 10 February 2018, the AFHSB’s study included 2,926 cases and 2,557 controls, with 89% and 90% having been vaccinated, respec-tively. After adjustment, VE for active com-ponent service members was statistically significant at 19% (95% CI, 3%–33%). For influenza A(H3N2) and influenza B, the adjusted VE estimates were 27% (95% CI, -9%–50%) and 25% (95% CI, -8%–48%), respectively (Figure); neither adjusted VE estimate was statistically significant.

E D I T O R I A L C O M M E N T

Overall, adjusted VE estimates for DoD studies were moderately protective for the dependent population. The AFHSB-AF sat-ellite’s overall adjusted VE was statistically significant and conferred moderate to high

protection; NHRC’s adjusted VE was statis-tically significant overall and was moder-ately protective for influenza A(H3N2); and AFHSB’s active component adjusted VE was statistically significant overall and provided some protection.

All of the VE studies had limitations. For example, specimens were obtained from those seeking care at a medical treatment facility or meeting the influenza-like illness case definition; therefore, less severe cases that did not seek medical attention were not included in the analyses. Individuals included in the DoD studies were younger than the general population, so VE could not be analyzed for older, higher-risk pop-ulations. Active component members are a highly immunized population, which may have a negative impact on VE estimates due to methodologic validity (i.e., limited unvac-cinated controls) and biologic effects (i.e., repeated vaccination). Lower sample size could have contributed to the reduction of statistical power in some DoD analyses.

The Centers for Disease Control and Prevention (CDC) reported lower VE at

36% (95% CI, 27%–44%), compared with all DoD studies with a dependent popu-lation. The CDC’s adjusted VE for influ-enza A(H3N2) was low at 25% (95% CI, 13%–36%), 67% (95% CI, 54%–76%) for influenza A(H1N1)pdm09, and 42% (95% CI, 25%–56%) for influenza B.3 Midseason results for the CDC did not closely match DoD midseason VE estimates. This differ-ence in VE estimates may be due, at least in part, to differences in the types of influ-enza vaccine used in DoD and in civilian populations. More than half of the influ-enza vaccine purchased and administered by the DoD was derived from cell culture propagation rather than from egg propa-gation.3 A rapid decline of VE for the vac-cine component influenza A(H3N2) that was egg-propagated has been seen in the past few years.4 Zost et al. reported that the current circulating influenza A(H3N2) viruses possess a new glycosylation site in antigenic site B of the hemagglutinin, and that the current egg-adapted A(H3N2) component of the vaccine does not have this mutation, which is hypothesized to

T A B L E . Department of Defense midseason influenza vaccine effectiveness (VE) estimates, 2017–2018

Population Cases Controlsa

Influenza type No. of cases

% vaccinated

No. of controls

% vaccinated

Crude VE (%) 95% CI Adjusted

VE (%)b 95% CI

Dependents (AFHSB-AF)

Overall 1,160 16 1,383 25 36 25–45 51 41–59

Influenza A(H3N2) 610 12 1,383 32 21 5–35 37 22–49

Influenza A(H1N1) 153 2 1,383 42 71 56–81 79 67–86Influenza B 390 8 1,383 36 39 23–52 60 49–70

Border civilians (NHRC)

Overall 201 13 114 24 52 13–74 55 17–75

Influenza A(H3N2) 156 13 114 24 50 6–73 52 9–75

Influenza B 41 12 114 24 55 -25–84 63 -5–87

Active component service members (AFHSB)

Overall 2,926 89 2,557 90 9 -8–24 19 3–33

Influenza A 2,539 89 2,557 90 9 -9–24 19 2–33

Influenza A(H3N2) 301 89 2,557 90 15 -25–42 27 -9–50

Influenza B 383 89 2,557 90 12 -25–37 25 -8–48CI, confidence interval; AFHSB-AF, Armed Forces Health Surveillance Branch-Air Force satellite cell; NHRC, Naval Health Research Center; AFHSB, Armed Forces Health Surveillance BranchaAll studies used unmatched, influenza test-negative controls.bAFHSB-AF adjusted for age group, month of illness, and region; NHRC adjusted for age group; and AFHSB adjusted for sex, age group, month of illness, and 5-year prior vaccination status (Y/N).

MSMR Vol. 25 No. 6 June 2018 Page 28

diminish antigenicity.5 Additional research is needed to assess whether VE against cir-culating A(H3N2) viruses varies by vaccine propagation type.

Author affiliations: Defense Health Agency/Armed Forces Health Surveillance Branch Air Force Satellite, U.S. Air Force School of Aero-space Medicine, Wright-Patterson AFB, OH (Ms. Shoubaki, Lt Col Federinko); Armed Forces Health Surveillance Branch, Silver Spring, MD (Dr. Eick-Cost, Ms. Hu, Ms. Lynch); Naval Health Research Center, San Diego, CA (Dr. Myers, Mr. Hawksworth).

Acknowledgments: The authors thank the DoD Global Respiratory Pathogen Surveil-lance Program and its sentinel site partners, the Navy and Marine Corps Public Health Center, and the Centers for Disease Con-trol and Prevention Border Infectious Dis-ease Surveillance Program in San Diego

and Imperial Counties, CA, which collected samples and case data from participating outpatient clinics.

Disclaimer: Dr. Myers is an employee of the U.S. Government. This work was prepared as part of his official duties. Title 17, U.S.C. §105 provides that copyright protection under this title is not available for any work of the U.S. Government. Title 17, U.S.C. §101 defines a U.S. Government work as work prepared by a military service member or employee of the U.S. Government as part of that person’s offi-cial duties.

Report No. 18-XX was supported by Armed Forces Health Surveillance Branch under work unit no. 60805. The views expressed in this article are those of the authors and do not necessarily reflect the official policy or posi-tion of the Department of the Navy, Depart-ment of Defense, or the U.S. Government.

R E F E R E N C E S

1. Sullivan SG, Chilver MB, Carville KS, et al. Low interim influenza vaccine effectiveness, Aus-tralia, 1 May to 24 September 2017. Euro Surveill. 2017;22(43):17-00707.2. Cost AA, Hiser MJ, Hu Z, et al. Brief report: mid-season influenza vaccine effectiveness es-timates for the 2013–2014 influenza season. MSMR. 2014;21(6):15–17.3. Flannery B, Chung JR, Belongia EA, et al. In-terim estimates of 2017–18 seasonal influenza vac-cine effectiveness—United States, February 2018. MMWR Morb Mortal Wkly Rep. 2018;67(6):180–185.4. Wu NC, Zost SJ, Thompson AJ, et al. A struc-tural explanation for the low effectiveness of the seasonal influenza H3N2 vaccine. PLoS Pathog. 2017;13(10):e1006682.5. Zost SJ, Parkhouse K, Gumina ME, et al. Con-temporary H3N2 influenza viruses have a glycosyl-ation site that alters binding of antibodies elicited by egg-adapted vaccine strains. Proc Natl Acad Sci USA. 2017;114(47):12578–12583.

F I G U R E . Department of Defense midseason influenza vaccine effectiveness (VE), 2017–2018

-200 -150 -100 -50 0 50 100 150VE (%)

Overall, AFHSB-AF - Dependents (1,160; 1,383)

Flu type, study site, population (no. of cases; no. of controls)

Overall, AFHSB - Active component service members (2,926; 2,557)

A(H3N2), AFHSB-AF - Dependents (610; 1,383)

A(H3N2), NHRC - Border civilians (156; 114)

Influenza B, AFHSB-AF - Dependents (390; 1,383)

79% (67–86)

60% (49–70)

27% (-9–50)

52% (9–75)

37% (22–49)

19% (3–33)

Adj. VE (95% CI)

Influenza B, NHRC - Border civilians (41; 114)

A(H3N2), AFHSB - Active component service members (301; 2,557)

A(H1N1), AFHSB-AF - Dependents (153; 1,383)

63% (-5–87)

25% (-8–48)Influenza B, AFHSB - Active component service members (383; 2,557)

51% (41–59)

Overall, NHRC - Border civilians (201; 114)55% (17–75)

CI, confidence interval; AFHSB-AF, Armed Forces Health Surveillance Branch–Air Force Satellite; AFHSB, Armed Forces Health Surveillance Branch; NHRC, Naval Health Research Center

June 2018 Vol. 25 No. 6 MSMR Page 29

MSMR’s Invitation to ReadersMedical Surveillance Monthly Report (MSMR) invites readers to submit topics for consideration as the basis for future MSMR reports. The

MSMR editorial staff will review suggested topics for feasibility and compatibility with the journal’s health surveillance goals. As is the case with most of the analyses and reports produced by Armed Forces Health Surveillance Branch staff, studies that would take advantage of the healthcare and personnel data contained in the Defense Medical Surveillance System (DMSS) would be the most plausible types. For each promising topic, Armed Forces Health Surveillance Branch staff members will design and carry out the data analysis, interpret the results, and write a manuscript to report on the study. This invitation represents a willingness to consider good ideas from anyone who shares the MSMR’s objective to publish evidence-based reports on subjects relevant to the health, safety, and well-being of military service members and other beneficiaries of the Military Health System (MHS).

In addition, the MSMR encourages the submission for publication of reports on evidence-based estimates of the incidence, distribution, impact, or trends of illness and injuries among members of the U.S. Armed Forces and other beneficiaries of the MHS. Information about manu-script submissions is available at www.health.mil/MSMRInstructions.

Please email your article ideas and suggestions to the MSMR Editor at dha.ncr.health-surv.mbx.msmr@mail.mil.

Letter to the Editor

To the Editor: As both a formermember of the Armed ForcesEpidemiological Board and the

Defense Health Board and as an investiga-tor who has studied the prevalence of hepa-titis C virus (HCV) among recruits, I was interested to read the brief report by Taylor and colleagues in the December 2017 issue of the MSMR.1

It was a little surprising to read the authors’ statement that “. . . the prevalence among military recruits accessioning into the U.S. Air Force has not been described.” May I respectfully remind the authors that in 2000, a manuscript titled “45-Year Fol-low-up of Hepatitis C Virus Infection in Healthy Young Adults” was published in the Annals of Internal Medicine? The manu-script reported studies of serum samples for hepatitis C antibody from more than 8,000 military recruits at the Fort Francis E. War-ren Air Force Base obtained between 1948 and 1954.2 Those authors described 0.2% of the tested recruits as having positive HCV studies by ELISA and by recombinant immunoblot assay. Because these sera were drawn upon entrance into the Air Force,

one could not determine how many of this population subsequently contracted HCV either while on active duty or after their military career. However, of the 17 recruits in the latter report with a positive antibody test for HCV, only one (5.9%) had died of liver disease 42–45 years after the original phlebotomy. Such data are required from larger and more recent cohorts to com-plete this important natural history evalua-tion even if there are newer (and expensive) therapeutic approaches to individuals with hepatitis C infections.

The reported cohort represents an opportunity for the authors of the brief report to identify and to carry out a mean-ingful comparison between the two cohorts and also to plan/attempt long-term follow-up for their recently identified cohort as the resulting data would have both prac-tical and public health implications for the Services, and especially for Veterans Administration healthcare programs. Data from identified cohorts of military person-nel can provide important information for the future. This aspect was recently empha-sized in a special supplement of Military

Medicine in October 2015 and is important for more than only hepatitis infections.3

I hope that the authors can take these factors into consideration.

Edward L. Kaplan, MD

Author affiliation: Professor Emeritus, Department of Pediatrics, University of Min-nesota Medical School, Minneapolis, MN

R E F E R E N C E S

1. Taylor DF, Cho RS, Okulicz JF, Webber BJ,Gancayco JG. Brief report: Prevalence of hepati-tis B and C virus infections in U.S. Air Force basicmilitary trainees who donated blood, 2013–2016.MSMR. 2017;24(12):20–22.2. Seeff LB, Miller RN, Rabkin CS, et al. 45-yearfollow-up of hepatitis C virus infection in healthyyoung adults. Ann Intern Med. 2000;132(2):105–111.3. Lindler LE, Gaydos JC (eds.). Military Medicine(Supplement). Special Issue: Assessing exposuresusing emerging laboratory technologies and biore-pository specimens. Mil Med. 2015;180(10).

MSMR Vol. 25 No. 6 June 2018 Page 30

June 2018 Vol. 25 No. 6 MSMR Page 31

Medical Surveillance Monthly Report (MSMR)Armed Forces Health Surveillance Branch11800 Tech Road, Suite 220 Silver Spring, MD 20904

MEDICAL SURVEILLANCE MONTHLY REPORT (MSMR), in continuous publication since 1995, is produced by the Armed Forces Health Surveillance Branch (AFHSB). The MSMR provides evidence-based estimates of the incidence, distribution, impact, and trends of illness and injuries among U.S. military members and associated populations. Most reports in the MSMR are based on summaries of medical administrative data that are routinely provided to the AFHSB and integrated into the Defense Medical Surveillance System for health surveillance purposes.

Archive: Past issues of the MSMR are available as downloadable PDF files at www. health.mil/MSMRArchives.

Online Subscriptions: Submit subscription requests at www.health.mil/MSMRSubscribe.

Editorial Inquiries: Call (301) 319-3240 or send email to: dha.ncr.health-surv.mbx. msmr@mail.mil.

Instructions for Authors: Information about article submissions is provided at www. health.mil/MSMRInstructions.

All material in the MSMR is in the public domain and may be used and reprinted without permission. Citation formats are available at www.health.mil/MSMR.

Opinions and assertions expressed in the MSMR should not be construed as reflecting official views, policies, or positions of the Department of Defense or the United States Government.

Follow us: www.facebook.com/AFHSBPAGE

http://twitter.com/AFHSBPAGE

ISSN 2158-0111 (print) ISSN 2152-8217 (online)

Chief, Armed Forces Health Surveillance Branch COL Douglas A. Badzik, MD, MPH (USA)

EditorFrancis L. O’Donnell, MD, MPH

Contributing EditorsLeslie L. Clark, PhD, MSShauna Stahlman, PhD, MPH

Writer/EditorValerie F. Williams, MA, MS

Managing/Production EditorElizabeth J. Lohr, MA

Data AnalysisStephen B. Taubman, PhDSaixia Ying, PhD

Layout/DesignDarrell Olson

Editorial OversightCol Dana J. Dane, DVM, MPH (USAF) COL P. Ann Loveless, MD, MS (USA)CDR Shawn S. Clausen, MD, MPH (USN)Mark V. Rubertone, MD, MPH